Applying blood-derived epigenetic algorithms to saliva: cross-tissue similarity of DNA-methylation indices of aging, physiology, and cognition
Abstract Background Epigenetic algorithms of aging, health, and cognition, based on DNA-methylation (DNAm) patterns, are prominent tools for measuring biological age and have been linked to age-related diseases, cognitive decline, and mortality. While most of these methylation profile scores (MPSs)...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-04-01
|
| Series: | Clinical Epigenetics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s13148-025-01868-2 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Background Epigenetic algorithms of aging, health, and cognition, based on DNA-methylation (DNAm) patterns, are prominent tools for measuring biological age and have been linked to age-related diseases, cognitive decline, and mortality. While most of these methylation profile scores (MPSs) are developed in blood tissue, there is growing interest in using less invasive tissues like saliva. The aim of the current study is to probe the cross-tissue intraclass correlation coefficients (ICCs) of MPSs developed in blood applied to saliva DNAm from the same people. While our primary focus is on MPSs that were previously found to be robustly correlated with social determinants of health, including second- and third-generation clocks and MPSs of physiology and cognition, we also report ICC values for first-generation clocks to enable comparison across metrics. We pooled three publicly available datasets that had both saliva and blood DNAm from the same individuals (total n = 107, aged 5–74 years), corrected MPSs for cell composition within each tissue, and computed the cross-tissue ICCs. Results We found that after correcting for cell composition, saliva–blood cross-tissue ICCs were moderate for second- and third-generation indices of aging and MPSs of physiology and cognition. Specifically, PCGrimAge had the highest ICC (0.76), followed by PCPhenoAge (0.72), a measure of cognitive performance (Epigenetic-g, 0.69), DunedinPACE (0.68), PCGrimAge Acceleration (0.67), PCPhenoAge Acceleration (0.66), an MPS of hs-CRP (0.58), and BMI (0.54). These ICCs appear lower than previous reports on within-tissue ICCs (saliva ICCs range from 0.67 to 0.85, blood ICCs range from 0.73 to 0.93). Cross-tissue ICCs values for first-generation biological age acceleration measures were poor, ranging from 0.19 to 0.25. Conclusions Our findings suggest that applying second- and third-generation MPSs of biological age acceleration and related phenotypes developed in blood to saliva DNAm results in moderate cross-tissue similarity and the precise cross-tissue correspondence differs by measure. While the degree of cross-tissue similarity of several MPSs may suffice for some research settings, it may not be suitable in clinical or commercial applications. Collection of both blood and saliva DNAm samples is necessary to validate existing algorithms and to customize MPSs in saliva DNAm. |
|---|---|
| ISSN: | 1868-7083 |