- Open Access
Erratum to: DNA methylation age of human tissues and cell types
© Horvath; licensee BioMed Central. 2015
- Received: 19 March 2015
- Accepted: 19 March 2015
- Published: 13 May 2015
The original article was published in Genome Biology 2013 14:3156
Fortunately, all of the other statements about cancer remain intact since the coding error effectively added an offset term to predicted age that changed little with chronological age (Figure 1). I am comforted by the fact that most of the reported results for cancer become even more significant, including the following. First, the results for cancer tissues are now more congruent with those obtained for cancer cell lines (which remain unchanged). Second, the age predictor leads to a much lower error in cancer tissues (now 16 years). Third, the results for TP53 become more significant, that is TP53 mutations are associated with lower age acceleration in colorectal cancer.
Figure seven in the original publication; Figure 2 here: Age acceleration versus number of somatic mutations in the TCGA data.
Figure eight in the original publication; Figure 3 here: Age acceleration in breast cancer.
Figure nine in the original publication; Figure 4 here: Age acceleration in colorectal cancer, glioblastoma multiforme and acute myeloid leukemia.
Additional file twelve in the original publication; Additional file 1 here: Description of cancer data sets.
Additional file thirteen in the original publication; Additional file 2 here: DNAm versus chronological age in cancer.
Additional file fourteen in the original publication; Additional file 3 here: Age acceleration versus tumor grade and stage.
Additional file fifteen in the original publication; Additional file 4 here: Age acceleration versus mutation count status in breast cancer.
Additional file sixteen in the original publication; Additional file 5 here: Selected significant gene mutations versus age acceleration.
Additional file seventeen in the original publication; Additional file 6 here: Effect of TP53 mutation on age acceleration.
Below, for sections of the original paper that are affected by the error, I explain how the corrected results are different from those that were reported.
In the original paper, I reported the correlation between DNAm age and chronological age as being 0.15 (P = 1.0×10−29). The correct correlation is 0.16 (p = 2.5×10−33; Additional file 2A). In addition, I reported that each cancer/affected tissue shows evidence of significant age acceleration. Instead, out of 20 cancer/affected tissues, only 6 exhibit positive age acceleration effects while others often show negative age acceleration effects, i.e. they appeared younger than expected (Figure 2B).
The following original statement remains unchanged: “Tumor morphology (grade and stage) has only a weak relationship with age acceleration in most cancers: only 4 out of 33 hypothesis tests led to a nominally (p < 0.05) significant result (Additional file 3)”.
But I have to retract the statement that only the negative correlation between stage and age acceleration in thyroid cancer remains significant after a applying a Bonferroni correction. It turns out that the uncorrected p-value of 0.0048 (Additional file 3Z) is not significant after multiplying it by 33.
Cancer tissues with high age acceleration exhibit fewer somatic mutations
The original statement that the number of mutations per cancer sample tends to be inversely correlated with age acceleration (Figure 2A) remains unchanged.
But I have to retract the claim that one can observe a significant negative relationship between age acceleration and the number of somatic mutations in thyroid cancer (Figure 2O).
TP53 mutations are associated with lower age acceleration
Additional file 5 presents the genes whose mutation has the strongest effect on age acceleration. The following original statement remains unchanged: "Strikingly, TP53 was among the top 2 most significant genes in 4 out of the 13 cancer data sets".
But I have to revise the following paragraph:
"Further, TP53 mutation is associated with significantly lower age acceleration in five different cancer types including AML (p = 0.0023), breast cancer (p = 1.4E-5 and p = 3.7E-8), ovarian cancer (p = 0.03), and uterine corpus endometrioid (p = 0.00093). Further, marginally significant result can be observed in lung squamous cell carcinoma and colorectal cancer (p = 0.073, below). I could only find one cancer type (GBM) where mutations in TP53 are associated with a nominally significant increased age acceleration (p = 0.02)".
TP53 mutation is associated with significantly lower age acceleration in six data sets (Additional file 6) including AML (P = 0.0041), breast cancer (P = 7.8×10−12 and P = 1.4×10−12), ovarian cancer (P = 0.04), uterine corpus endometrioid (P = 0.0012), and colorectal cancer (P = 0.036, Figure 4B). Further, marginally significant result can be observed in lung squamous cell carcinoma (P = 0.088 Additional file 6G).
Somatic mutations in steroid receptors accelerate DNAm age in breast cancer
The following original statement remains unchanged: “Age acceleration differs greatly across different breast cancer types (Figure 4N): Luminal A tumors (typically ER+ or PR+, HER2-, low Ki67), show the highest positive age acceleration”.
But I retract the statement that luminal B tumors (typically ER+ or PR+, HER2+ or HER2- with high Ki67) show a similar effect.
Proto-oncogenes affect DNAm age in colorectal cancer
The p-value in the following statement "Echoing previous results, TP53 mutations appear to be associated with decreased age acceleration (p = 0.073)" needs to be revised to (p = 0.036, Figure 4B).
The p-value in the following statement "Promoter hypermethylation of the mismatch repair gene MLH1 leads to the most significant increase in age acceleration (P = 5.7×10−5)" needs to be revised to (p = 3.9×10−7, Figure 4D).
The p-value in the following statement "The CpG island methylator phenotype, defined by exceptionally high cancer-specific DNA hypermethylation, is also significantly (p = 3.5×10−5) associated with age acceleration" needs to be revised to (p = 3.6×10−7, Figure 4F).
DNAm age in glioblastoma multiforme (GBM)
The p-value in the following statement "Interestingly, age acceleration in GBM samples is highly significantly (p = 3.3×10−7) associated with certain mutations in H3F3A" needs to be revised to (p = 0.0015, Figure 4J).
The p-value in the following statement "…age acceleration varies significantly (p = 2×10−7) across the GBM subtypes defined in (Sturm et al 2012)" needs to be revised to (p = 1.2×10−7, Figure 4L).
Acute myeloid leukemia
The following statement remains unchanged: “Mutations in FLT3, RAS, NPMc, and various well characterized translocations do not seem to relate to age acceleration in AML samples”.
But I have to retract the claim that mutations in IDH1 do not relate to age acceleration. Rather, IDH1 mutations are nominally significantly related with age acceleration (p = 0.036, Figure 4S).
DNAm age of cancer cell lines
My original results for cancer cell lines were not affected by the coding error;that is, they remain correct.
My conclusion section remains largely unchanged. But I have to revise the following sentence: "While all cancer tissues exhibit signs of severe age acceleration, this is not necessarily the case for individual cancer cell lines".
It turns out that cancer types are similar to individual cancer cell lines. Some cancer types exhibit positive age acceleration effects (e.g. luminal breast cancer) while others exhibit negative age acceleration (e.g. basal breast cancer, Figure 3N).
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0) which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.