On 2025-05-06 21:54:54, user Young Cho wrote:
Key findings
The first primary discovery discussed in this paper is that age affects DNA methylation (DNAm) in different ways depending on the location in the genome. For example, the researchers found that promoter regions experienced hypermethylation with age, while transposable elements experienced hypomethylation. The second discovery was that female dogs had lower methylation in X chromosome-linked long interspersed nuclear element-1s (LINE1s) than male dogs. This was unexpected because X chromosome inactivation would normally lead to much higher methylation, and thus a possible explanation is that females are more susceptible to age-related decline in methylation than males. The third discovery is that size is largely associated with the rate of methylation decline, where larger dogs lose LINE1 methylation with age at a significantly higher rate than smaller dogs. This may be a large part of the reason why larger dogs tend to have shorter lifespans.
Results
Figure 1 shows how the researchers prepared, processed, and annotated the blood samples from 864 dogs in order to measure the DNAm at various CpG sites throughout their genome.
Figure 2 shows how age affects DNAm at different locations in the dogs’ genomes.
Figure 3 shows that X-linked and age-associated X-linked LINE1s are more methylated in male dogs compared to female dogs.
Figure 4 plots LINE1 methylation against age, showing that large dogs lose LINE1 methylation faster and have shorter lifespans than small dogs.
These figures effectively show each significant conclusion the researchers made. I liked that only the major figures were included in the main part of the paper, while other less important, supporting data was labeled as supplementary figures.
Supplemental figure 1 shows that there was no significant difference in the sizes of the female and male dogs used in this study, and also that each dog was appropriately categorized by size so that there was minimal overlap.
Supplemental figure 2 gives an overview of the CpG sites that were analyzed in this study, showing CpG densities, region lengths, mean methylation, and high sample coverage.
Supplemental figure 3 shows that promoter regions that are affected by methylation as one ages are often associated with important biological functions, such as integrated stress response signaling, immune response, etc.
Discussion
The discussion clearly outlines the main findings of this study. The first of which, correlates age with methylation in the genome. Additionally the methylation of LINE1s, a class of autonomous transposable elements, had a lower frequency in females than males. Lastly, the rate of DNA methylation in LINE1 was also associated with dog size. These conclusions were also supported by previous studies in other model species. For instance, the loss of methylation of LINE1s in other mammals was also associated with age, an observation also made in dogs. Furthermore, the authors conveyed the importance of this study (i.e. the higher frequency of LINE1 methylation in males) through linking the loss of silencing in Y-linked LINE1s with shorter life span, as this was observed in human studies.
Methods
In the methods section, describing blood sample collection, mention of transporting blood samples between universities was included. However the temperature at which these samples were at during this time period was not recorded. Additionally the amount of sample taken was not mentioned. Rather, this section conveyed what cohort of dogs their samples were taken from, who isolated specific cells from the samples, and why these peripheral blood mononuclear cells were isolated. Additionally, the method to extract the peripheral blood mononuclear cells may be present in reference 115. At the time of writing this review, this reference could not be accessed.
Sample annotation, describes how information about the dogs’ backgrounds were obtained, first through survey. Which was utilized to categorize the dogs by size. Following this, a machine learning model calculated adult dog heights from 0-4. Where 0 equated to ankle height and 4 was waist high. For this, it may be preferable to designate height values or ranges rather than descriptions, as these heights are subjective. However, utilizing calculated or predicted heights allowed several advantages compared to survey or self-reported data, which was also included.
RRBS was used to identify CpG sites by converting unmethylated cytosines to uracils, which is highly appropriate for this study, because it allowed the researchers to find and analyze any age-related changes in methylation across their samples. These sites were grouped into 47,393 regions based on their distances from each other, such that CpG sites from the same region were no further than 250 bp apart. This was an important preprocessing step because CpG sites that are close to each other are likely to experience similar levels of methylation, and thus any sites located further could have negatively skewed their data. Furthermore, a large majority of CpG regions (80%) had greater than 5x coverage in more than 95% of samples, ensuring that the data they collected was consistent and therefore reliable.
The researchers annotated their CpG regions by using previously made data as a reference. For instance, they used NIH CanFam4, which is the current reference genome for Canis lupus familiaris, to align their sequenced data with. They were also able to map chromatin states (e.g. active promoters, enhancers, heterochromatin, and inactive regions) using Son et al.’s epigenetic map.
PQLSeq was used for region-based differential methylation analysis because, as RRBS provides methylation and total read counts for CpG sites, PQLSeq uses binomial proportions to model the methylation data as an age-related factor and account for fixed and random effects of sex, predicted height, and genetic relatedness.
With a generally large sample size of 864 dogs, each with their own larger set of methylated regions, it was necessary to use the Sol Supercomputer in order to compile and process all of the data efficiently and accurately.
Strengths and limitations
This paper was enlightening. There was a clear reason as to why this research should continue as these results can shed light on epigenetic processes that occur as mammals age. The goal of this paper was also achieved, as the authors recorded that DNA methylation plays a role as dogs age, where specific DNA regions experience changes in methylation frequency. The methods utilized to obtain these results and the following data supported their conclusions. Wherein their discussion section, ample prior results in various model species supported their results. Furthermore, the authors stated some limitations in their study but were able to come to conclusions utilizing papers focusing on other species. Thus, analysis of their results were sufficient.
However, there are areas for improvement. There are minor adjustments that may need to be considered before publishing, which were detailed in the editorial decision. Furthermore, in terms of the cohort utilized, the number of female to male subjects and breed sizes varied between these categories and was not addressed. Which may inspire the question of how many subjects would be considered sufficient for the conclusions made in this paper. Additionally, figure clarity can be improved, specifically for Figure 4. As in print, the color differentiating breed size are very similar, to increase visibility it may be preferable to utilize other colors or dotted/dashed lines. However, these are minor changes and inquiries that do not diminish the importance of this paper.
Editorial decision
There are some formatting issues. For instance, in the third paragraph of the discussion section in the fourth sentence, stating, “A genetic mechanism underlying this breed’s increased cancer risk. . .” between, “breed’s,” and, “increased,” is a double space. Additionally, in the third paragraph of the introduction section, DNA methylation’s acronym (DNAm) is defined for the second time, which can be omitted, as the acronym was fully spelled out in a previous paragraph. The references are not formatted using current APA format. Furthermore there are some inconsistencies with the addition of DOIs, as only a couple references have them. In references that name species, they are not italicized. Additionally reference 115, at the time of writing this review, was not accessible. More clarity would also be preferable regarding how 864 subjects were chosen out of the approximate 1,000 dogs enrolled in the cohort used. Furthermore, during sample annotation, height values categorized by a machine learning model were described rather than providing a range of measurements, as the descriptions were subjective (e.g. “ankle high”). Otherwise, the value of this study is clear and this project is certainly valuable to understanding the impact of methylation and aging in mammals. Overall, the paper was a great read.
This paper was organized into sections in a way that made it very easy to follow. We especially liked that major conclusion statements were bolded and made as subheadings in the discussion section, and that only figures that directly supported those conclusions were included in the main part of the paper. One other minor thing that we would say could be improved upon is the golden retriever case study — in order to show that age-related methylation patterns affect dogs on a breed-specific level, we feel like it would have been interesting to see the comparison of methylation in golden retrievers to distinct breeds rather than just non-golden retrievers. Overall, I would accept this paper with very minor revisions.