Table 1 Median [IQR] of each variance component and the corresponding median [IQR] of the intraclass correlation coefficient (ICC) across CpG sites for study duplicates and technical replicates, by DNA source, with and without normalisation using ComBat.
Type of duplicate | Type of sample | ComBat | Subject | Study | Plate | Chip | Residual | ICC |
|---|---|---|---|---|---|---|---|---|
Technical replicates | PBMC | No | 45% [17–72%] | 2% [1–5%] | 3% [1–7%] | 5% [2–8%] | 36% [17–62%] | 0.55 [0.22–0.81] |
Yes | 41% [13–69%] | 0.7% [0.3–1.4%] | 0.0% [0.0–0.2%] | 0.0% [0.0–0.1%] | 57% [30–85%] | 0.42 [0.14–0.70] | ||
Technical replicates | DBS | No | 21% [4–45%] | 2% [1–4%] | 3% [2–7%] | 3% [1–5%] | 63% [44–79%] | 0.25 [0.05–0.50] |
Yes | 20% [5%–41%] | 0.3% [0.0–0.7%] | 0.0% [0.0–0.0%] | 0.0% [0.0–0.0%] | 79% [59–94%] | 0.20 [0.05–0.41] | ||
Study duplicates | PBMC | No | 24% [6–53%] | 2% [1–5%] | 3% [1–7%] | 5% [3–8%] | 56% [36–74%] | 0.29 [0.8–0.59] |
Yes | 27% [8–54%] | 0.7% [0.3–1.4%] | 0.0% [0.0–0.2%] | 0.0% [0.0–0.1%] | 71% [45–90%] | 0.27 [0.08–0.54] | ||
Study duplicates | DBS | No | 11% [3–35%] | 2% [1–4%] | 3% [2–7%] | 3% [1–5%] | 71% [53–83%] | 0.13 [0.03–0.38] |
Yes | 12% [3–34%] | 0.3% [0.0–0.7%] | 0.0% [0.0–0.0%] | 0.0% [0.0–0.0%] | 87% [65–96%] | 0.12 [0.03–0.34] |
