DNA Methylation Age of Human Tissues and Cell Types

• p. 3 Section 'Predictive accuracy across different tissues', last paragraph

  brain data (Figures 1F and 2F,G)

• p. 5 Figure 2 panel F title

  F Brain Cerebellar err=5.9 cor=0.92, p=9.5e-09

Only the authors can say what is right here.

Judge's note: Zoomed render of Figure 2 row 2 from source/main.pdf: panel F title reads 'F Brain Cerebellar err=5.9 cor=0.92, p=9.5e-09', with samples labelled 54 up to about age 60. The legend gives the test-set error as 3.6 years. The other cited panels have small errors (1F 1.4, 2G 1.5, 2I 0.37). So 2F does not illustrate 'particularly accurate' by the paper's own error measure, though its correlation is high. Only the author can say whether 2F was meant.

Second check: another model agreed

• p. 13 Conclusions

  Mutations in TP53 are associated with lower DNAm age.

Judge's note: Read the Conclusions on PDF page 13 and the results section headed "TP53 mutations are associated with lower age acceleration". The results speak only of age acceleration, in five cancer types, and note that in glioblastoma TP53 mutation goes with increased age acceleration (P = 0.02). The conclusion sentence is loose shorthand and drops the GBM exception; it is imprecise wording rather than a wrong result.

Second check: another model agreed

• p. 3 Section 'Predictive accuracy across different tissues'

  not be used for studying whether DNAm is well cali

Judge's note: Read the crop of the 'Predictive accuracy across different tissues' paragraph. The sentence is about the age estimate, and a few lines later the same paragraph writes 'whether DNAm age is poorly calibrated'. Missing word, meaning recoverable.

Second check: none (judge only)

• p. 4 Figure 1 legend (F)

  pre-frontal cortex (cor = 0.98, 1.4)

• p. 5 Figure 2 legend (S)

  uterine endometrium (cor = 0.55, 11)

• p. 4 Figure 1 caption (F)

  (F) pre-frontal cortex (cor = 0.98, 1.4)

Judge's note: Read both legends on the page images; every other panel entry writes 'error = x'. The panel titles confirm the bare numbers are the errors (1F err=1.4, 2S err=11). Cosmetic.

Second check: none (judge only)

• p. 5 Figure 2 legend (T)

  10 Epstein Barr Virus transformed B cell

• p. 4 DNAm age and progeria

  Epstein-Barr virus-transformed B cells

Judge's note: Both forms read on the page images (Figure 2 legend and the 'DNAm age and progeria' paragraph). Hyphenation and capitalisation are inconsistent; style only.

Second check: none (judge only)

• p. 18 Abbreviations

  LOO: Leave one data set out

• p. 5 Left column

  set-out cross-validation (LOOCV) analysis

• p. 12 Section 'Acute myeloid leukemia'

  Mutations in IDH1 (similar to the case of GBM), FLT,

• p. 14 Figure 9, panel T title

  AML, FLT3 p = 0.15

Judge's note: Checked each against images and text. Page 12 crop prints "FLT," in italics; the Figure 9 panel T title and x-axis label on the page 14 image read FLT3. Page 17 image prints "H3K27m3"; the Methods text has "H3K27me3". A text search finds LOOCV on page 5 and in the Figure 3 and Additional file 1 legends, and "LOO" alone only in the abbreviation list. All three are cosmetic naming inconsistencies; the gene name FLT is the most worth fixing.

Second check: none (judge only)

• p. 6 Left column, coefficient of variation paragraph

  jejunum, penis, pancreas, esophagus, spleen, pancreas, lymph node, diaphragm

• p. 6 Section 'DNAm age of multiple tissues from the same subject'

  esophagus, spleen, pancreas, lymph node, diaphragm

• p. 6 DNAm age of multiple tissues from the same subject

  esophagus, spleen, pancreas, lymph node

Judge's note: I checked the crop of page 6, left column. 'pancreas' appears twice in the same list of example tissues. One of the two is a duplicated word, or a different tissue was meant. This is cosmetic.

Second check: another model agreed

• p. 11 TP53 mutations section

  marginally significant result can be observed in lung squamous cell carcinoma

Judge's note: Read on the page image (PDF p11); the sentence is printed as quoted. Two cancers are named, and the Additional file 17 description uses 'Marginally significant results'. Grammar only.

Second check: none (judge only)

• p. 13 Left column, cell lines paragraph

  varies greatly across the cancer lines lines

• p. 18 Additional file 18 description

  lines lines (y-axis)

• p. 13

  varies greatly across the cancer lines lines (Additional files

• p. 13

  ies greatly across the cancer lines lines (Additional files 1

Judge's note: The body-text occurrence is visible on the page image (PDF p13, left column, 'varies greatly across the cancer lines lines'). The second occurrence in the Additional file 18 description (PDF p18) was read from the text layer, where both words sit in running text on one line.

Second check: another model agreed, code check

• p. 14 Figure 9, panel T title

  21 T AML, FLT3 p = 0.15

Judge's note: Rendered the Figure 9 bottom row from source/main.pdf at 5x zoom. Panel titles read "S AML, IDH1 p = 0.22", "21 T AML, FLT3 p = 0.15", "U AML, RAS p = 0.82". Only panel T has a number before the letter; it is part of the figure itself.

Second check: none (judge only)

• p. 14 Figure 9 legend

  Promoter hyper methylation of the mismatch repair gene MLH1

Judge's note: Page image shows "hyper methylation" at the start of a legend line with a clear space, so it is not a hyphenation artefact. The body text has "hypermethylation" as one word throughout, including the matching sentence "promoter hypermethylation of the mismatch repair gene MLH1".

Second check: none (judge only)

• p. 15 Short description of the healthy tissue data sets

  42 and 43 consist of involve whole blood samples

• p. 15 same section

  Data set 19 normal consists of

• p. 15 same section

  sets 27 consists of normal adjacent tissue from TCGA

• p. 15

  Data set 19 normal consists of adjacent colon tissue

Judge's note: All three phrases appear on the page 15 crops as quoted: two verbs in a row, a misplaced 'normal', and plural 'sets' with a single data set and a singular verb. The meaning can still be recovered, so this is cosmetic.

Second check: another model agreed

• p. 15 Short description of the healthy tissue data sets, data set 45

  Simple Collection [46]

Only the authors can say what is right here.

Judge's note: Rests on outside knowledge: that the autism cohort resource is named the Simons Simplex Collection. The page image of PDF page 15 does print "Simons Simple Collection [46]", across a line break with no hyphenation involved. Nothing in the article or reference 46 (Alisch et al., Genome Res 2012) gives the cohort name, so the correct name cannot be confirmed from the files.

Second check: none (judge only)

• p. 15 Materials and methods, data set 45

  Simons Simple Collection

• p. 15

  healthy children of the Simons Simple Collection

Only the authors can say what is right here.

Judge's note: Rests on outside knowledge: the autism cohort is named the Simons Simplex Collection. The page 15 image does read "Simons Simple Collection". Nothing in the files states the cohort's correct name.

Second check: none (judge only)

• p. 15 Materials and methods, data set 39

  Data set 39 consists of thyroid TCGA (THCA data)

Judge's note: Reads as quoted on the page image of PDF page 15. Every neighbouring TCGA entry uses "from TCGA" (data sets 19, 24, 26, 27, 29, 30, 35, 38), so a word has dropped out. Cosmetic; the meaning is clear.

Second check: none (judge only)

• p. 16 Additional file 4 description

  correlation in panel 0) reflects very young subjects

• p. 16 Additional file 4 description

  (B-S) Results for individual blood data sets.

• p. 16

  negligible age correlation in panel 0) reflects very young

• p. 16 Additional file 4 legend

  The negligible age correlation in panel 0) reflects

Judge's note: The crop shows "panel 0)" as printed, with no opening bracket; panels are lettered A to T, so the zero is a mis-set letter O. Panel S is described separately as a bar plot across cord blood data sets, so the individual-data-set range should presumably end at R. Both are cosmetic slips in the legend; the supplement itself is not in the files, so the intended letter cannot be checked against the figure.

Second check: another model agreed

• p. 16 Additional file 4 description

  The negligible age correlation in panel 0) reflects

Judge's note: Viewed the crop: the glyph is the narrow digit zero and the opening bracket is missing, and the text layer also has the digit 0. The surrounding legend lists panels (A), (B-S), (S), (T), so a lettered panel O within B-S is the evident intent. Same legend also uses (S) twice (within B-S and for the cord blood bar plot), which is not part of this cluster.

Second check: none (judge only)

• p. 17

  Negatively related CpG are often located chromatin state 4 regions

• p. 17 Additional file 9 legend

  more likely to be in chromatin states 2 (weak promoters)

Judge's note: Read on the page 17 image; all three slips are present as quoted. Purely grammatical; meaning is clear.

Second check: another model agreed

• p. 17 Additional file 7 legend

  hereditary disorders (50 genes, 2.6E-4)

Judge's note: Read on the page 17 image. The inconsistent formatting within one list is present as quoted; cosmetic.

Second check: none (judge only)

• p. 17 Additional file 7 legend

  cell death and survival (P = E-37)

Only the authors can say what is right here.

Judge's note: Read on the page 17 image; "P = E-37" appears twice in the same sentence. Every other P-value in the legend has a mantissa (1.1E-7, 3.7E-5, 9.2E-5, 2.6E-4). The intended value (probably 1E-37) cannot be determined from the files, so it is a query to the author.

Second check: none (judge only)

• p. 18 Reference 3

  Biokhimiia 1967, 31:88–993

• p. 18 Reference 4

  mehylcytosine

• p. 18 Reference 11

  Mugatroyd C

• p. 19 Reference 53

  JL: Rego-Pérez I. Blanco FJ:

Only the authors can say what is right here.

Judge's note: Rests on outside knowledge: that the first author of this paper is Murgatroyd, not Mugatroyd. The page 18 image prints "Mugatroyd C" as quoted. This cluster bundles four reference-list items; the other three are adjudicated in their own clusters (ref 3 page range in C056, "mehylcytosine" in C055, ref 53 author list in C058), so this verdict covers only the reference 11 surname.

Second check: another model agreed

• p. 18

  HUVEC: Cell- human umbilical vascular endothelial cells;

• p. 18 Abbreviations

  HUVEC: Cell- human umbilical vascular endothelial cells

• p. 18 Abbreviations

  LOO: Leave one data set out

Judge's note: Read on the page 18 image; the stray "Cell-" is printed as quoted. The reviewer's further point that HUVEC stands for "vein" rather than "vascular" rests on outside knowledge and is left as a possible query. The LOO/LOOCV point in this cluster is covered by C052.

Second check: another model agreed

• p. 18 Additional file 18 legend

  across various cancer lines lines (y-axis)

• p. 18

  bserved across various cancer lines lines (y-axis). The DNAm

Judge's note: Read on the page 18 image: "cancer lines" ends one line and "lines (y-axis)" begins the next, so the word is printed twice. Probably "cancer cell lines" was intended.

Second check: code check

• p. 18 Reference 4

  The 5 mehylcytosine in DNA of rats

Judge's note: Read on the page 18 image; the title is printed "The 5 mehylcytosine in DNA of rats", missing the "t". A plain misspelling with an evident fix, not an extraction artefact.

Second check: none (judge only)

• p. 19 Reference 53

  Fernández JL: Rego-Pérez I. Blanco FJ:

Judge's note: Read on the page 19 image. The punctuation is printed as quoted, and "Rego-Pérez I." is set in italics like a journal name, so the author list is broken mid-way. All other entries separate authors with commas and use a single colon before the title.

Second check: none (judge only)