Massively Parallel Digital Transcriptional Profiling of Single Cells
2017 · paper · Biomedicine
Describes a droplet-based system that counts 3' mRNA in tens of thousands of single cells per sample, with about 50% cell capture. It was tested on cell lines, 68k blood cells and transplant bone marrow, where it determined host and donor chimerism.
Checked: PDF from nature.com, fetched 4 Oct 2026. Page numbers are those of that PDF.
12 findings
Substance "clusters 1, 4 and 6 are CD8+ cytotoxic T cells, whereas clusters 2 and 3 are CD4+ T cells" contradicts the following text ("cluster 2 represents naive CD8 T cells, whereas cluster 4 represents naive CD4 T cells (Fig. 3c)") and the Fig. 3c labels "2: Naive CD8+" and "4: Naive CD4+"; clusters 2 and 4 are swapped in the first sentence.
- p. 5 Results, Subpopulation discovery, right column
clusters 1, 4 and 6 are CD8 þ cytotoxic T cells, whereas clusters 2 and 3 are CD4 þ T cells
- p. 5 Results, last lines of page
suggests that cluster 2 represents naive CD8 T cells, whereas
- p. 6 Results, first line
cluster 4 represents naive CD4 T cells (Fig. 3c)
- p. 5 Figure 3c, cluster assignment labels
2: Naive CD8+
Judge's note: Read the paragraph on the page 5 image and its continuation on page 6, and the Fig. 3c left-hand labels (6: CD8+, 1: Activated CD8+, 3: Memory and Reg T, 2: Naive CD8+, 4: Naive CD4+). All refer to the same 10 k-means clusters of Fig. 3b; no other numbering is introduced. The sentence should read clusters 1, 2 and 6 CD8+ and clusters 3 and 4 CD4+. (One reviewer note about 'clusters 1-3 and 6' belongs to C009.)
Second check: another model agreed
Report a problem with this finding
Substance ">80% T cells (enrichment of CD3D, part of the T-cell receptor complex, in clusters 1–3 and 6)": by Fig. 3b these clusters total 70.0%; cluster 4 (11.2%, "Naive CD4+") is needed to reach the >80% (Methods: T is "81%").
- p. 5 Results, Subpopulation discovery, right column
complex, in clusters 1–3 and 6)
- p. 5 Figure 3b cluster percentages
(11.2%)
- p. 11 Methods, Identification of cluster-specific genes
size of B, dendritic and T is 5.7%, 6.6% and 81%, respectively
- p. 5 Results, Subpopulation discovery, right column
80% T cells (enrichment of CD3D, part of the T-cell receptor complex, in clusters 1–3 and 6)
Judge's note: Fig. 3b percentages read from the page 5 image: 1 (9.3%), 2 (31.2%), 3 (17.0%), 4 (11.2%), 6 (12.5%). Clusters 1, 2, 3, 6 sum to 70.0%; adding cluster 4 gives 81.2%, matching the Methods' 81% on page 11. The same paragraph and Fig. 3c treat cluster 4 as a T-cell cluster, and no statement excludes it. The list should read clusters 1–4 and 6.
Second check: another model agreed · checked by hand
Report a problem with this finding
Substance Results: "identify minor genotypes as low as 3% at >95% sensitivity and PPV (Fig. 4a,b). A minor population could not be detected when the mixed ratio was below 3% (Fig. 4c)"; Methods: "the level of detection of the method, which in in silico mixing experiments was determined to be 4% of 6,000 cells".
- p. 6 Results, Genotype-based method, right column
identify minor genotypes as low as 3% at 495% sensitivity
- p. 6 Results, Genotype-based method
when the mixed ratio was below 3% (Fig. 4c)
- p. 11 Methods, SNV-based genotype assignment
mixing experiments was determined to be 4% of 6,000 cells
- p. 6 Results, Genotype-based method, right column
A minor population could not be detected when the mixed ratio was below 3% (Fig. 4c)
Only the authors can say what is right here.
Judge's note: Read p6 right column and the p11 Methods paragraph on the page images. Both describe the same in silico B/C mixing of 6,000 cells. Results and Fig. 4c (confident calls at 3% and above, plus one confident call at 2%) give 3% as the detection limit; Methods give 4%. No statement in the article reconciles the two (the 90%-of-cells posterior >75% rule is the criterion that produces Fig. 4c's confident/not-confident calls, so it should yield the same limit). Only the authors can say which figure is right.
Second check: another model agreed
Report a problem with this finding
Reference or label Figure 2f axis labels "293T SNV counts" (y) and "Jurkat SNV counts" (x) are interchanged relative to the data: the 293T-only sample lies along the "Jurkat SNV counts" axis and the Jurkat-only sample along the "293T SNV counts" axis.
- p. 4 Figure 2f, y-axis label
293T SNV counts
- p. 4 Figure 2f, x-axis label
Jurkat SNV counts
- p. 4 Figure 2 legend (f)
293T- and Jurkat-enriched SNVs were plotted for each sample
Judge's note: Looked at Fig. 2f on the page 4 image and crop. In the '293T-only' panel the blue (293T-only) cells spread to about 150 on the x-axis labelled 'Jurkat SNV counts' with about 0 on the y-axis; in the 'Jurkat-only' panel the red cells rise along the y-axis labelled '293T SNV counts'. The 1:99 293T:Jurkat panel has its dominant red population on the '293T SNV counts' axis too, so the panel titles and colours are consistent with each other and it is the axis labels that are swapped. The legend gives no alternative definition.
Second check: another model agreed
Report a problem with this finding
Reference or label Fig. 3 legend: "(d–i) tSNE projection of 68k PBMCs, with each cell coloured based on their normalized expression of CD3D, CD8A, NKG7, FCER1A, CD16 and S100A8", but panels g–i show only the three sub-clusters of myeloid cluster 9 ("1 (37%)", "2 (47%)", "3 (16%)").
- p. 5 Fig. 3 caption (d-i)
tSNE projection of 68k PBMCs, with each cell coloured based on their normalized expression of CD3D, CD8A, NKG7, FCER1A, CD16 and S100A8
- p. 5 Fig. 3g-i panels
1 (37%)
Judge's note: Viewed Fig. 3 on the page 5 image and crop: panels d–f show the full 68k tSNE map, whereas g–i (FCER1A, CD16, S100A8) show a different, smaller embedding with three clusters labelled 37%, 47% and 16%. The text on page 6 says k-means was applied to cluster 9 cells and cites Fig. 3g–i for dendritic cells, CD16+ and CD16-/low monocytes. The legend does not mention this and reuses cluster numbers 1–3 that clash with panel b.
Second check: another model agreed
Report a problem with this finding
Reference or label Figure 4 title "Genotype analysis of in silico and in vitro mixing of PBMCs." but panels a, b and c are all described as in silico mixing; no in vitro mixing data are shown in the figure (those are in Table 1).
- p. 7 Figure 4 title
Genotype analysis of in silico and in vitro mixing of PBMCs.
- p. 7 Figure 4 legend, panel c
Called mix fraction versus actual mix fraction in in silico mixing of PBMCs from Donors B and C
- p. 7
on versus actual mix fraction in in silico mixing of PB
Judge's note: Read the Figure 4 legend on the p7 image and crop. All three panel descriptions say in silico mixed PBMCs from Donors B and C. The experimentally mixed samples (50:50, 90:10, 99:1) appear only in Table 1. The 'repeated in' flagged by the text rule ("in in silico") is correct English with the italic Latin term and is not a defect.
Second check: another model agreed, code check
Report a problem with this finding
Reference or label ddPCR primer list is mis-set: "AGC/ZEN/CACGATGATACAT/3IABkFQ/OR. AOV1_f, 5′-CCGGAAGTGGGTCTCGTOR-3′ and AOV1_r, 5′-TTCTTCATAGCCTTCCCGATACCOR-3′; AOV1_p": a stray "OR" follows the AP1S3 probe and ends both primer sequences, and the gene is printed as "AOV1".
- p. 10 Methods, ddPCR assay, probe sequences
AGC/ZEN/CACGATGATACAT/3IABkFQ/OR. AOV1_f
- p. 10 Methods, ddPCR assay, probe sequences
GGTCTCGTOR-30 and AOV1_r
- p. 10 Methods, ddPCR assay, probe sequences
TTCTTCATAGCCTTCCCGATACCOR-30
Judge's note: Confirmed on the enlarged p10 crop; the stray 'OR' is in the typeset page. 'O' is not a nucleotide code, and the three stray 'OR' strings sit exactly where the gene-name prefix would precede each italic 'AOV1', so the 'OR' evidently belongs to the gene name (ORAOV1; the full gene name is from outside knowledge, but the defect and the fix of removing 'OR' from the sequences do not depend on it). As printed, the two primer sequences are wrong.
Second check: another model agreed
Report a problem with this finding
Wording Results: "The capture rate is consistent across four types of cells with cell loads ranging from ~1,000 to ~23,000 (Supplementary Table 2)"; Discussion: "load from 1,000 to 23,000 cells per well, from four different cell lines and two primary cell types (PBMCs and BMMCs)"; Methods name only three cell lines ("Jurkat (ATCC TIB-152), 293T (ATCC CRL-11268) and 3T3 (ATCC CRL-1658)").
- p. 2 Results, Technical demonstration
across four types of cells with cell loads ranging from B1,000 to
- p. 9 Discussion, second paragraph
from four different cell lines and two
- p. 9 Methods, Cell lines and transplant patient samples
Jurkat (ATCC TIB-152), 293T
- p. 9 Discussion, paragraph 2
from four different cell lines and two primary cell types (PBMCs and BMMCs)
Only the authors can say what is right here.
Judge's note: Checked page 2 and page 9 images. The same loading range is attributed to four cell types in the Results and to six (four lines plus two primary types) in the Discussion, and the article describes only three cell lines. Supplementary Table 2 is not among the files, so a fourth cell line cannot be excluded, but the Results/Discussion counts would still differ. Authors must say which is right.
Second check: another model agreed
Report a problem with this finding
Wording Text claims "minor genotypes as low as 3% at >95% sensitivity and PPV (Fig. 4a,b)", but in Fig. 4b the PPV at 3% minor population is about 0.93 when the major population is Donor B (it passes 0.95 only at 4%); the Donor B curves also show a call at 2% (PPV about 0.87), against "could not be detected when the mixed ratio was below 3%".
- p. 6 Results, in silico mixing
as low as 3% at >95% sensitivity
- p. 7 Fig. 4b
Positive predictive value (PPV) versus percentage of minor
Only the authors can say what is right here.
Judge's note: Rendered Fig. 4b from the PDF at 6x. Red (major population B) points: 2% about 0.87, 3% about 0.93, 4% about 0.95; blue (major C) at 3% about 0.97. So the >95% PPV claim at 3% holds for one direction only. Values are read from an untabulated curve, hence author query; the gap at 3% (about 0.93 vs 0.95) is clear at this magnification. Related to C016 but a separate defect (PPV claim vs figure).
Second check: another model agreed
Report a problem with this finding
Wording Garbled temperature in Methods, Cell preparation: "FBS + 10% DMSO, freezing to − by °C in a CoolCell® FTS30 (BioCision)"; the temperature value is missing.
- p. 9 Methods, Cell preparation
FBS þ 10% DMSO, freezing to by C in a CoolCells FTS30
- p. 9 Methods, Cell preparation
freezing to
Only the authors can say what is right here.
Judge's note: Confirmed on the enlarged crop: the typeset page prints '− by °C', so this is not an extraction artefact. The value is presumably −80 °C as in the other freezing steps, but only the authors can confirm. The same sentence also has the odd "frozen at 10×" (probably a lost cell-concentration unit). Rated S3 rather than cosmetic because a protocol value is lost.
Second check: another model agreed
Report a problem with this finding
Wording SNV rule for the Jurkat/293T mixture says "the fraction of counts from the other species is <0.2. Cells with a fraction of SNV from either species between 0.2 and 0.8 are considered multiplets", but Jurkat and 293T are both human cell lines; 'species' should be 'cell line'.
- p. 10 Methods, SNV analysis of Jurkat and 293T scRNA-seq data
from the other species is o0.2. Cells with a fraction of SNV from either
Judge's note: Read the 'SNV analysis of Jurkat and 293T scRNA-seq data' paragraph on the page 10 image. The paragraph concerns only Jurkat and 293T (both human; the paper itself calls them human cell lines in the Results). No second species is involved, so the wording is carried over from the human/mouse experiment. The meaning is recoverable, hence wording-level.
Second check: another model agreed
Report a problem with this finding
Wording "For cell classification comparison between purified and frozen PBMCs, we pooled all the cells labelled as T or natural killer cells together": 'purified' should be 'fresh'.
- p. 11 Methods, Comparison between fresh and frozen PBMCs
For cell classification comparison between purified and frozen PBMCs
Judge's note: Read on the page 11 image (top of right column). The section is headed 'Comparison between fresh and frozen PBMCs', opens with '68k fresh PBMCs and 3k frozen PBMCs', and the same paragraph ends 'the comparison between fresh and frozen cells'. Purified populations are the classification reference, not one of the two samples being compared, so 'purified' is a slip.
Second check: another model agreed
Report a problem with this finding
21 minor (typos and the like; most checked by the judge only)
Wording "we estimated that 1,012 GEMs contained cells, of which 482 and 538 contained reads that mapped primarily to the human and mouse transcriptome, respectively": 482 + 538 = 1,020, eight more than 1,012.
- p. 2 Results, Technical demonstration
we estimated that 1,012 GEMs
- p. 2 Results, Technical demonstration
482
- p. 2 Results, Technical demonstration
538
- p. 2 Results, technical demonstration
482 and 538
Only the authors can say what is right here.
Judge's note: Page 2 image confirms the numbers; Fig. 2a (page 4) prints "1,012 cells recovered, Multiplet rate=1.6%". The excess of 8 equals the "Eight cell-containing GEMs" with substantial human and mouse counts in the next sentences, so the multiplets are very probably counted in both species groups; the Methods (page 10) define the multiplet rate as twice the rate of mixed barcodes, and 2*8/1012 = 1.58% matches the printed 1.6%. The data are therefore likely sound, but the paper nowhere says the eight are counted in both groups and "of which ... primarily" reads as a partition. Wording issue, not a wrong value; proposals such as 530 are not supported.
Second check: another model agreed · checked by hand
Report a problem with this finding
Wording Methods for the Jurkat/293T PCA say "The normalized UMI counts of each gene are used to show expression of a marker in a tSNE plot", but the article shows this experiment's marker (CD3D) only on a PCA plot (Fig. 2e: "PC1 and PC3 are plotted, and each cell is colored by the normalized expression of CD3D").
- p. 10 Methods, PCA analysis of mixing of Jurkat and 293T cells
marker in a tSNE plot.
- p. 4 Figure 2 legend (e)
PC1 and PC3 are plotted, and each cell is colored by the normalized expression of
Only the authors can say what is right here.
Judge's note: Read the Methods paragraph on page 10 and the Fig. 2 legend on page 4. Searched the article text for tSNE: every tSNE mention concerns PBMCs or BMMCs; none concerns the Jurkat/293T mixture. The sentence looks carried over from the PBMC section. The supplement (Supplementary Fig. 4) is not among the files, so a tSNE of these samples there cannot be excluded; hence a query rather than a correction.
Second check: another model agreed
Report a problem with this finding
Wording "which are absent in AML035 post-transplant sample": Fig. 5b shows a thin Immature Granulocytes (purple) segment of about 1% in the AML035 post-transplant (donor) bar, and a few purple points appear in that tSNE panel.
- p. 8 Results, right column
which are absent in AML035 post-transplant sample
- p. 8 Figure 5b, bar 'AML035 post-transplant (donor)'
AML035 post-transplant (donor)
Judge's note: Rendered Fig. 5b from the PDF at 6x: the AML035 post-transplant bar has a purple sliver at about 86-87% of the axis (roughly 1% of cells), next to a thin blue one. The Fig. 5a crop shows a small purple group in the AML035 post-transplant panel. 'Absent' is a slight overstatement of 'about 1%'; the contrast with about 18-20% in AML027 host cells stands.
Second check: another model agreed
Report a problem with this finding
Wording "Since the relative cluster size of B, dendritic and T is 5.7%, 6.6% and 81%, respectively, we would expect cluster 10 ... to contain multiplets consisting mostly from B:dendritic (0.36%) and B:dendritic:T (0.3%)": the product of the stated B and dendritic sizes is 0.38%, not 0.36%.
- p. 11 Methods, Identification of cluster-specific genes
B:dendritic (0.36%) and B:dendritic:T (0.3%)
- p. 11 Methods, Identification of cluster-specific genes
5.7%, 6.6% and 81%
Judge's note: Numbers read on the page 11 image. The triple product reproduces the printed 0.3%, which indicates the figures are simple products of cluster fractions; the pair product then gives 0.376%. Rounding of the inputs does not reach 0.36% (the lowest values that round to 5.7% and 6.6% give 0.370%). The Results point to Supplementary Methods for this calculation, which are not in the files, so a different formula or unrounded inputs cannot be excluded. The gap is 0.02 percentage points and does not affect the argument.
Second check: another model agreed
Report a problem with this finding
Typo Author contributions never mention J.G.U. ("Jason G. Underwood1,3" in the author list), and Junjie Zhu is abbreviated both "J.J.Z." (software sentence, competing interests) and "J.Z." ("G.X.Y.Z., P.R., T.S.M., J.Z., K.R.L. and M.T.G. analysed the data").
- p. 1 Author list
Jason G. Underwood1,3
- p. 12 Author contributions
G.X.Y.Z., J.M.T., P.B., P.R., Z.W.B., T.S.M., B.J.H., J.H.B., E.A.S. and J.P.R. designed
- p. 12 Competing financial interests
T.S.M. and B.J.H. are employees of 10x Genomics.
- p. 12 Author contributions, last two sentences
T.S.M., J.Z., K.R.L. and M.T.G.
Only the authors can say what is right here.
Judge's note: Read the author list on page 1 and the Author contributions and Competing financial interests on page 12. All other authors' initials appear in the contributions; J.G.U. appears nowhere on page 12, including the list of 10x Genomics employees although affiliation 1 is 10x Genomics (he also has affiliation 3, so the competing-interest omission may be legitimate). The only author who could be J.Z./J.J.Z. is Junjie Zhu, so the two forms are an inconsistency. Only the authors can say what J.G.U. contributed.
Second check: none (judge only)
Report a problem with this finding
Typo Fig. 2 legend mixes spellings in one sentence ("are colored in red and termed 'Human only'; and cell barcodes with significant mouse and human reads are coloured in grey"), and the pipeline is "Cell Ranger" in the Introduction and Methods but "CellRanger" in the Results ("processed by CellRanger") and Fig. 1f legend.
- p. 4 Figure 2 legend (a)
are colored in red and termed ‘Human only’; and cell barcodes with significant mouse and human reads are coloured
- p. 2 Results, Technical demonstration
by CellRanger (Supplementary Methods and Fig. 1f)
- p. 2 Introduction
Ranger,
Judge's note: Both inconsistencies seen on the page 4 legend and page 2 text images (Introduction 'Cell Ranger', Results 'CellRanger'). Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo "~7% myeloid cells (enrichment of S100A8 and S100A9 (ref. 20) in cluster 9 (Supplementary Methods, Fig. 3b–f, Supplementary Fig. 5e and Supplementary Data 3).": the parenthesis opened at "(enrichment" is never closed.
- p. 5 Results, right column
S100A8 and S100A9 (ref. 20) in cluster 9 (Supplementary
Judge's note: Counted parentheses on the page 5 image: the three preceding parallel items close theirs ("...in clusters 1–3 and 6)", "...in cluster 5)", "...in cluster 7)"), but the myeloid item has no closing parenthesis after "cluster 9".
Second check: none (judge only)
Report a problem with this finding
Typo Author name misspelled in Methods: "similar to the approach used by Macoscko et al.7"; reference 7 and the Results spell it "Macosko".
- p. 11 Methods, PCA and tSNE analysis of PBMCs
approach used by Macoscko et al.7
- p. 11 Methods, PCA and tSNE analysis of PBMCs
Macoscko et al.7
- p. 12 Reference 7
Macosko, E. Z. et al.
- p. 5 Results text and reference 7
Macosko et al.7
Judge's note: 'Macoscko' read on the page 11 image; reference 7 on the page 12 image reads 'Macosko, E. Z. et al.'; the Results text (page 5) reads 'Macosko et al.7'. Same reference number, so the same author.
Second check: none (judge only)
Report a problem with this finding
Typo Missing word: "expected to be present a PBMC sample." should read "present in a PBMC sample".
- p. 6 Results, left column, first paragraph
expected to be present a PBMC sample.
- p. 6 End of first paragraph, left column
expected to be present a PBMC sample
Judge's note: Seen on the p6 page image, end of first paragraph, left column. The preposition is missing in the typeset page.
Second check: none (judge only)
Report a problem with this finding
Typo Table 2 footnote: "AML, acute myelod leukaemia." should be "myeloid".
- p. 7 Table 2 footnote
AML, acute myelod leukaemia.
- p. 7 Table 2 footnote
acute myelod leukaemia
Judge's note: Seen on the p7 page image under Table 2.
Second check: none (judge only)
Report a problem with this finding
Typo Misspelling: "Ninty-seven per cent is within the observed range" should be "Ninety-seven".
- p. 7 Results, right column
Ninty-seven per cent is within the observed
- p. 7 Transplant section
Ninty-seven per cent
Judge's note: Seen on the p7 page image, right column near the bottom.
Second check: none (judge only)
Report a problem with this finding
Typo Human gene symbols are written "GATA1" in the Results ("CD34+, GATA1+"; "CD34, GATA1, CD71 and HBA1") but "Gata2" and "Gata1" in the Methods ("and Gata2, a marker for early erythroids40").
- p. 8 Results, right column
cells and immature erythroids (CD34 þ , GATA1 þ )
- p. 11 Methods, PCA and tSNE analysis of BMMCs
and Gata2, a marker for early erythroids40
Judge's note: Checked p8 text and the p11 Methods crop. The capitalisation inconsistency is real. The alleged marker mismatch is not established: the Results phrase "blast cells and immature erythroids (CD34+, GATA1+)" can cover both 'Blasts and Immature Ery 1' (CD34, Gata2) and 'Immature Ery 2' (Gata1), which together dominate the AML027 post-transplant host bar in Fig. 5b (about 54% and 17%).
Second check: none (judge only)
Report a problem with this finding
Typo Discussion: "are encapsulated into GEMS at ~80% fill rate"; the plural is "GEMs" everywhere else in the article.
- p. 9 Discussion, paragraph 2
are encapsulated into GEMS at
- p. 9 Discussion, second paragraph
encapsulated into GEMS at
Judge's note: Capital S confirmed on the p9 crop. Every other occurrence in the text (about 20) is 'GEMs'.
Second check: none (judge only)
Report a problem with this finding
Typo Garbled phrase: "All of the known listed of barcodes that are 1-Hamming-distance away from an observed barcode are considered." ('listed of' for 'list of' or 'listed').
- p. 10 Methods, Alignment, barcode assignment and UMI counting
All of the known listed of
- p. 10 Methods, Alignment, barcode assignment and UMI counting
All of the known listed of barcodes
Judge's note: Visible on the page 10 image and the C035-a crop (right column), exactly as quoted. Not an extraction artefact.
Second check: none (judge only)
Report a problem with this finding
Typo Missing full stop (and dangling '×'): "Amplified cDNA product was cleaned up with the SPRIselect Reagent Kit (0.8 × ) cDNA was subsequently sheared to ~200 bp using a Covaris M220 system".
- p. 10 Methods, ERCC assay
cDNA was subsequently sheared to B200 bp using a Covaris M220
Judge's note: Seen in the ERCC assay paragraph on the page 10 image and enlarged crop. The parallel sentence in the library-construction section reads '(0.6 × SPRI). The cDNA was subsequently sheared', so a full stop (and 'SPRI') is missing here.
Second check: none (judge only)
Report a problem with this finding
Typo Product name capitalised inconsistently: "cleaned up with the SPRISelect Reagent Kit (0.8 × SPRI)" in the ERCC assay versus "SPRIselect Reagent Kit" elsewhere (including four lines later in the same paragraph).
- p. 10 ERCC assay
SPRISelect Reagent Kit (0.8
- p. 10 Sequencing library construction
SPRIselect Reagent Kit (0.6
Judge's note: Confirmed on the enlarged crop: capital S in 'SPRISelect' in the ERCC paragraph; 'SPRIselect' later in that paragraph and in the library-construction section. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Missing word: "'B' refers clusters 6 and 9 because of markers such as CD19 and CD79A" ('refers to').
- p. 11 Methods, PCA and tSNE analysis of BMMCs
‘B’ refers
- p. 11 Methods, PCA and tSNE analysis of BMMCs
‘B’ refers clusters 6 and 9
Judge's note: Read on the page 11 image. The neighbouring sentences read "'T' refers to cluster 10" and "'Mature Ery' refers to cluster 2".
Second check: none (judge only)
Report a problem with this finding
Typo "as per Stephens et al.38" but reference 38 lists a single author: "Stephens, M. Dealing with label switching in mixture models. J. R. Stat. Soc. Ser. B 62, 795–809 (2000)."
- p. 11 Methods, SNV-based genotype assignment
per Stephens et al.38
- p. 12 References, 38
Stephens, M. Dealing with label switching in mixture models.
Judge's note: In-text citation read on the page 11 image and entry 38 on the page 12 image. The entry as printed has one author and no 'et al.', so 'et al.' in the text contradicts the paper's own reference list. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo "Barnes-hut37 approximation to t-SNE": lower-case 'hut'; reference 37 itself prints "Barnes-Hut-SNE".
- p. 11 Methods, PCA and tSNE analysis of PBMCs
Barnes-hut37 approximation
Judge's note: Read on the page 11 image; reference 37 on the page 12 image has the capitalised form. Cosmetic capitalisation of an eponym, inconsistent with the paper's own reference list.
Second check: none (judge only)
Report a problem with this finding
Typo Author contributions give two forms for one author: "P.R., R.W., A.W., G.X.Y.Z., J.J.Z., T.S.M. and M.S.L. wrote the analysis software. G.X.Y.Z., P.R., T.S.M., J.Z., K.R.L. and M.T.G. analysed the data."
- p. 12 Author contributions
G.X.Y.Z., J.J.Z., T.S.M.
- p. 12 Author contributions
G.X.Y.Z., P.R., T.S.M., J.Z., K.R.L. and M.T.G.
Judge's note: Read on the page 12 crop. The author list on page 1 has only one author with initials J…Z (Junjie Zhu); the competing-interests statement also uses 'J.J.Z.'. So 'J.Z.' and 'J.J.Z.' denote the same person. Which form is right is trivial, but the two differ.
Second check: none (judge only)
Report a problem with this finding
Typo Reference 9 gives the year 2016 for a bioRxiv preprint whose own URL is dated 5 March 2014: "Preprint at http://biorxiv.org/content/early/2014/03/05/003236 (2016)."
- p. 12 Reference 9
http://biorxiv.org/content/early/2014/03/05/003236 (2016)
- p. 12 Reference 9
2014/03/05/003236 (2016)
Judge's note: Viewed the reference-list crop on PDF page 12; the entry reads exactly as quoted, so the URL date (2014/03/05) and the citation year (2016) disagree within the entry itself. No outside fact is needed for that. The other two preprints in the list carry their posting year, matching their identifiers (ref. 36 arXiv 1207.3907 "(2012)", ref. 37 arXiv:1301.3342 "(2013)"), so 2016 is out of line with the list's own practice for preprints. However, ref. 6, a Fluidigm web document, also carries "(2016)", so 2016 may be an access year applied to web-only sources; the year could also reflect a later revision. Either reading could be legitimate, hence an observation rather than a correction. The evident tidy-up would be "(2014)".
Second check: none (judge only)