Skip to main content
. 2022 Jan 14;11:e73420. doi: 10.7554/eLife.73420

Figure 2. Test-retest reliability of DunedinPACE.

The figure graphs DunedinPACE values for replicate Illumnina450k datasets for n = 36 individuals in the dataset published by Lehne and colleagues (Lehne et al., 2015) (GEO Accession GSE55763). The ICC for DunedinPACE in the Lehne dataset is 0.96 95% CI [0.92–0.98].

Figure 2.

Figure 2—figure supplement 1. Reliability of DunedinPACE, DunedinPoAm, and DNA methylation clocks.

Figure 2—figure supplement 1.

The figure plots intraclass correlation coefficients (ICCs) estimated from replicate DNA methylation datasets for DunedinPACE, original DunedinPoAm, and the DNA methylation clocks proposed by Horvath, Hannum et al., Levine et al. (PhenoAge), and Lu et al. (GrimAge). Error bars show 95% confidence intervals. Data are graphed from three datasets, 36 Illumina 450 k technical replicates from GSE55763 (Lehne et al., 2015), and 28 Illumina EPIC array technical replicates and 350 Illumina 450k-EPIC replicates from Sugden et al., 2020. The Sugden et al. EPIC technical replicate dataset (Sugden A) included arrays analyzed as part of the Dunedin Study dataset. The Sugden et al. 450k-EPIC dataset (Sugden B) was used to identify probes for inclusion in the machine learning analysis from which DunedinPACE was derived.