A Pseudo-Longitudinal Methylome Projection Framework Defines a Buccal PACE-like Aging-Rate Score from Cross-Sectional DNA Methylation Data
Authors:
Preprint
Abstract
Background: DNA methylation-based biomarkers have enabled robust estimation of biological age across tissues, and longitudinally trained measures such as DunedinPACE provide estimates of the pace of aging from blood methylomes. However, longitudinal methylation data are often unavailable, particularly for minimally invasive tissues such as buccal mucosa. Here, we developed a pseudo-longitudinal framework to estimate a buccal mucosa-derived PACE-like aging-rate score from cross-sectional methylome data. Methods: We used a buccal biological age estimator as an internal pseudo-time axis. Methylation beta-values were transformed to M-values, and CpG-specific smooth functions of biological age were fitted in cross-validation. Local derivatives of these functions were used to project each individual’s buccal methylome forward by a small time step. The projected methylome was converted back to beta-values, biological age was recalculated, and the change in biological age per unit time was defined as a pseudo-aging velocity. This raw velocity was transformed to a non-negative PACE-like score centered at 1.0. We then trained cross-fitted models to predict the derived score from buccal CpG methylation profiles. Results: In 151 individuals, the proposed score was reproducibly predicted from buccal methylomes in out-of-fold analysis, with a Pearson correlation of 0.706 and Spearman correlation of 0.710 between observed and predicted PACE-like scores. Sensitivity analyses across CpG selection size and regression models showed broadly consistent performance. In contrast, the proposed buccal PACE-like score showed only modest association with measured DunedinPACE, and alternative attempts to reconstruct DunedinPACE from buccal methylomes, including supervised proxy modeling and buccal-to-blood CpG imputation, showed limited sample-level performance. Conclusions: These results support the feasibility of deriving a tissue-specific PACE-like aging-rate score from cross-sectional buccal methylome data by treating biological age as a pseudo-time axis. The proposed score should not be interpreted as a replacement for blood-derived DunedinPACE, but rather as an exploratory buccal methylome dynamics index that may capture tissue-specific aging-related variation.Keywords
This publication has 14 references indexed in Scilit:
- DNA methylation aging clocks: challenges and recommendationsGenome Biology, 2019
- SeSAMe: reducing artifactual detection of DNA methylation by Infinium BeadChips in genomic deletionsNucleic Acids Research, 2018
- An epigenetic biomarker of aging for lifespan and healthspanAging, 2018
- DNA methylation-based biomarkers and the epigenetic clock theory of ageingNature Reviews Genetics, 2018
- The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cellsNature Biotechnology, 2014
- Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarraysBioinformatics, 2014
- DNA methylation age of human tissues and cell typesGenome Biology, 2013
- Low-level processing of Illumina Infinium DNA Methylation BeadArraysNucleic Acids Research, 2013
- Genome-wide Methylation Profiles Reveal Quantitative Views of Human Aging RatesMolecular Cell, 2013
- Comparison of Beta-value and M-value methods for quantifying methylation levels by microarray analysisBMC Bioinformatics, 2010