Analysis of variability and epigenetic age prediction across microarray and methylation sequencing technologies

Area of Science:

• Epigenetics
• Genomics
• Bioinformatics

Background:

• Epigenetic clocks are crucial for estimating biological age.
• Comparing different technologies for epigenetic clock application is essential for data interpretation.

Purpose of the Study:

• To compare technical variability and signal strength between Infinium MethylationEPIC v2.0 arrays and Twist Human Methylome Panel.
• To evaluate the robustness and reproducibility of principal component (PC) trained epigenetic clocks across these technologies.

Main Methods:

• Utilized 100 technical replicate samples from adult buccal cohorts.
• Compared methylation data from MethylationEPIC arrays and Twist methylation sequencing across 753,648 shared CpGs.
• Assessed four PC-trained epigenetic clocks (pcHorvath1, pcHorvath2, pcHannum, pcDNAm PhenoAge) for reproducibility using mean absolute replicate difference (MRD).

Main Results:

• Twist methylation sequencing exhibited skewed distributions and fewer correlated CpGs compared to MethylationEPIC arrays.
• MethylationEPIC arrays showed higher signal strength, while a subset of CpGs performed well on both platforms.
• PC-trained epigenetic clocks demonstrated robustness, with varying reproducibility depending on the clock and technology; non-PC versions showed significantly lower reproducibility.

Conclusions:

• Principal component training enhances epigenetic clock robustness and reproducibility across different methylation technologies.
• Technology-specific adjustments and careful clock selection are vital for accurate epigenetic age estimation.
• Differences in methylation data generation impact epigenetic clock performance, necessitating cautious interpretation.