Long-Read epigenetic clocks identify improved brain aging predictions.

Area of Science:

• Genomics
• Epigenetics
• Computational Biology

Background:

• Epigenetic clocks are crucial for estimating biological age.
• Current clocks often use array-based data from limited ancestries, impacting generalizability.
• There is a need for more accurate and inclusive aging biomarkers.

Purpose of the Study:

• To develop novel epigenetic clocks using long-read sequencing data.
• To create ancestry-aware aging models trained on diverse populations.
• To evaluate the performance of long-read methylation data in aging clock construction.

Main Methods:

• Utilized GenoML, an automated machine learning platform.
• Trained aging clocks on DNA methylation data from Oxford Nanopore long-read sequencing.
• Incorporated over 28 million CpG sites from individuals of African and European ancestry.

Main Results:

• Developed accurate epigenetic clocks leveraging long-read methylation data.
• Demonstrated improved generalizability of aging clocks across ancestries.
• Highlighted the utility of large-scale, multi-ancestry datasets for biomarker development.

Conclusions:

• Long-read sequencing data is powerful for building accurate, ancestry-aware epigenetic clocks.
• Inclusive training datasets are essential for robust and generalizable biological aging estimates.
• This work advances the development of precision aging biomarkers.