Comprehensive Integration of Single-Cell Data

• p. 45 Figure 3E legend

  datasets from the mouse visual cortex

• p. 8 Results

  3,482 cells from the prefrontal cortex

• p. 29 Methods, Transferring cell type labels onto scATAC-seq

  for the mouse prefrontal cortex and scRNA-seq cells from the mouse visual cortex

Judge's note: Page 29 reads 'scATAC-seq cells ... for the mouse prefrontal cortex and scRNA-seq cells from the mouse visual cortex'. The legend's plural 'datasets' attributes both datasets to visual cortex. This is a wording misattribution in the legend; the Methods are correct.

Second check: another model agreed

• p. 48 Figure 5D legend

  Predicted expression patterns for four genes not originally profiled by STARmap

• p. 12 Results

  We identified four representative cases (Figure 5D)

• p. 48 Figure 5D panel

  Tesc, Pvrl3, Sox10, Grm2, Tcrb

Only the authors can say what is right here.

Judge's note: Panel D on p48 shows five genes under 'Predicted'. Sox10 also appears in panel C (osmFISH), possibly as a comparison, but neither the legend nor the text explains the fifth tile. Only the authors can say which four are meant.

Second check: another model agreed

• p. 16 Parameters for Seurat v3 integration

  For all neuronal, bipolar, and pancreatic analyses we choose a dimensionality of 30.

• p. 32 STARmap imputation with Drop-seq data

  choosing a dimensionality of 50

Judge's note: Both statements are on the page images (p16, p32). This may be a preprocessing setting rather than the integration dimensionality, but the paper does not say so. It is a minor gap in the general parameter statement.

Second check: another model agreed

• p. 20 Anchor scoring

  anchors representing reference and cell pairs have significantly higher scores than incorrect anchors

• p. 20 Anchor scoring

  anchors representing reference and cell pairs have significantly higher scores

Judge's note: The page 20 image confirms the phrase. Figure 2I and its legend compare 'Correct' (same cell type) with 'Incorrect' anchors, so the phrase misdescribes the comparison. This is a wording issue only.

Second check: another model agreed

• p. 25 Analysis of CD69+ bone marrow population

  fitting a three-component mixture model

• p. 25

  by fitting a three-component mixture model

Only the authors can say what is right here.

Judge's note: I read p.25 (main text lines 1184-1189). The text says a three-component normalmixEM fit was used and the cells were then grouped into high and low expressers. The handling of the third component is not stated anywhere, so only the authors can resolve this.

Second check: another model agreed

• p. 1 Abstract, final sentences

  spatial gene expression patterns.Our work presents

• p. 1 Abstract

  spatial gene expression patterns.Our work

Judge's note: The page 1 image shows the words run together with no space after 'patterns.', so this is not an extraction artefact.

Second check: none (judge only)

• p. 13 Results

  [Moffit et al., 2018]

• p. 36 Reference list

  Moffit JR, Bambah-Mukku D

• p. 36 Reference list

  Moffit JR, Bambah-Mukku D, Eichhorn SW

• p. 3 Introduction

  [Moffitt et al., 2018]

Judge's note: On page 36 there are two consecutive entries with the same authors, title and year, one with the misspelled 'Moffit' and the incomplete 'Science 5324'. Page 13 shows the in-text 'Moffit'.

Second check: none (judge only)

• p. 8 Results

  MGE-derived populations (LHX6)

• p. 45 Figure 3F legend

  Lhx6 600

• p. 45 Figure 3 legend (F)

  Gad2, Pvalb, Sst, Vip, Lamp5, and Id2 1,000; Lhx6 600

• p. 44 Figure 3F

  Neurod6 Gad2 Lhx6 Vip Lamp5 Sst Pvalb

Judge's note: The page 8 image shows italic 'LHX6' next to 'Gad2', and the figure and legend use 'Lhx6'. This is a cosmetic inconsistency. Reviewer note 3's point about the Id2 track is handled under C004.

Second check: none (judge only)

• p. 10 Transferring continuous and multimodal data

  fill in missing modalites in key datasets

Judge's note: The page 10 image shows 'modalites'.

Second check: none (judge only)

• p. 37 Reference list

  R Testi JHP, Lanier LL, 1989

• p. 11 Results

  [R Testi and Lanier, 1989]

Judge's note: Rests on outside knowledge for the exact correct form (Testi R, Phillips JH, Lanier LL). That the printed author field 'R Testi JHP' is malformed can be seen in the image, and it produces the irregular in-text citation on p11.

Second check: none (judge only)

• p. 12 Results, spatial mapping

  not originally profiled profiled by STARmap

• p. 12 Results, spatial mapping

  were not originally profiled profiled by STARmap

• p. 20 Methods

  anchors representing reference and cell pairs

• p. 12

  enes that were not originally profiled profiled by STARmap. We iden

Judge's note: The page 12 image shows 'profiled profiled'. Reviewer note 3's point about 'reference and cell pairs' is handled under C015.

Second check: code check

• p. 15 Discussion, final paragraph

  including both transcriptional, posttranscriptional, and translational regulation

Judge's note: The page 15 image confirms the wording. This is cosmetic.

Second check: none (judge only)

• p. 16 Parameters for Seurat v3 integration

  fluctuate across across datasets

• p. 16

  One parameter we expect to fluctuate across across datasets

• p. 16

  ameter we expect to fluctuate across across datasets represents

Judge's note: The page 16 image shows 'across across'.

Second check: code check

• p. 29 Latent semantic indexing

  Cusanovich and Hill et at.

• p. 29 Latent semantic indexing

  We then computed to log of this TF-IDF matrix

• p. 24 CITE-seq cross-validation

  by it’s contribution

• p. 27 mnnCorrect

  we performed principle component analysis

Judge's note: I verified each phrase on the page images. All are cosmetic: principal, et al., the log, its.

Second check: none (judge only)

• p. 19 CCA methods

  obtained as the left and right singular values from the SVD

Judge's note: On page 19, Λ is defined as the diagonal of singular values (λ^1/2), and u_i = γ_i, v_i = δ_i are columns of the singular-vector matrices Γ and Δ. The word 'values' is a slip. The equations themselves are correct, so the result stands.

Second check: none (judge only)

• p. 21 Anchor weighting equation for W_c,i

  Wc, i =

Judge's note: I viewed the crop. The denominator shows Σ with subscript 1 and superscript 'j = k.weight'. The intended meaning (sum over j from 1 to k.weight) can still be recovered, so this is a cosmetic notation error.

Second check: none (judge only)

• p. 28 Mixing metric

  accurately reflected the the mixing

• p. 24 CITE-seq cross-validation

  by it’s contribution

• p. 28 Mixing metric

  reflected the the mixing

• p. 29 LSI

  We then computed to log of this TF-IDF matrix

Judge's note: I verified each phrase on the page images. The other items in this cluster ('it’s contribution', 'computed to log', 'et at.') are handled under C027.

Second check: code check

• p. 25 Validation of CD69+ T cell population

  (Giagen, cat #: 79216)

• p. 25

  RLT lysis buffer (Giagen, cat #: 79216)

Only the authors can say what is right here.

Judge's note: Rests on outside knowledge: the supplier of RLT buffer is Qiagen. The page image shows 'Giagen'.

Second check: none (judge only)

• p. 25 Validation of CD69+ T cell population

  AmPure bead cleanup

Only the authors can say what is right here.

Judge's note: Rests on outside knowledge: the product is branded 'AMPure'. The page image shows 'AmPure'. This is cosmetic.

Second check: none (judge only)

• p. 26 Identification of rare subtypes in the Tabula Muris dataset

  (FindAnchors and IntegrateData)

• p. 25 Protein expression transfer to the HCA

  FindIntegrationAnchors and IntegrateData functions

Judge's note: I viewed the crop. FindAnchors appears only here (grep). Everywhere else, including the anchor-identification and HCA sections, the function is FindIntegrationAnchors.

Second check: none (judge only)

• p. 52 Key resources table

  Wolock et al., 2018

• p. 25 Methods

  Scrublet [Wolock et al., 2019]

• p. 40 References

  Wolock SL, Lopez R, Klein AM, 4 2019

Judge's note: I checked the page 52 table ("Wolock et al., 2018"), the page 25 Methods text ("Scrublet [Wolock et al., 2019]"), and the page 40 reference list, which has only one Wolock entry, dated 2019. A search of the text found a second in-text citation, also 2019 (line ~1452). The reference list has no 2018 Wolock entry, so the table's year does not match any reference. This is a minor inconsistency in a citation year.

Second check: none (judge only)

• p. 30 Methods heading

  Psuedo-bulk ATAC-seq data collation

Judge's note: The page image shows the heading 'Psuedo-bulk'; the same paragraph uses 'pseudo-bulk'.

Second check: none (judge only)

• p. 32 Analysis of integrated STARmap data

  We then performed differential expressed to identify

• p. 32

  We then performed differential expressed to identify gene expression markers

Judge's note: The page image shows the wrong word form exactly as quoted.

Second check: none (judge only)

• p. 35 Reference list, Hie et al.

  data. bioRvix.

Judge's note: The image shows 'bioRvix' as printed. Elsewhere the reference list spells it 'biorxiv' (www.biorxiv.org).

Second check: none (judge only)

• p. 35 Reference list

  Shenhav1 R

• p. 35 Reference list, Halpern et al., 2017

  Halpern KB, Shenhav1 R

• p. 39 van Dijk reference

  doi.org/10.10167j.cell.2018.05.061

Judge's note: Both are visible in the page images at normal type size, not as superscripts, so they are printed and are not extraction artefacts. The other Cell DOIs in the list use the form 10.1016/j.cell.

Second check: none (judge only)

• p. 37 Reference list

  Mccarroll SA, Saunders A, Macosko EZ, Wysoker A

Judge's note: The image confirms that Saunders A, Macosko EZ and Wysoker A each appear twice.

Second check: none (judge only)

• p. 43 Figure 2K y-axis

  Silhoutte Metric

Judge's note: Seen on the full-page image p43, panel K.

Second check: none (judge only)

• p. 45 Figure 3F legend

  by pooling together cells with for each cell type

• p. 45 Figure 3 legend (F)

  pooling together cells with for each cell type

Judge's note: Printed on p45 as quoted. A word is missing or 'with' is left over from an edit.

Second check: none (judge only)

• p. 47 Figure 4E legend

  similar set of deferentially expressed genes

• p. 47 Figure 4 legend (E)

  deferentially expressed genes

• p. 45 Figure 3 legend (F)

  pooling together cells with for each cell type

Judge's note: Printed on p47 as quoted. The Fig. 3F phrase also mentioned by one reviewer is covered by C039.

Second check: none (judge only)

• p. 48 Figure 5 legend (C)

  a highly sensitive single molecular assay

Judge's note: The legend on p48 reads as quoted. 'single-molecule' is the usual term, and 'single molecular assay' can be misread as 'one assay'. This is a minor wording inconsistency.

Second check: none (judge only)