Enlarged leukocyte referent libraries can explain additional variance in blood-based epigenome-wide association...

Area of Science:

• Genomics and Epigenetics
• Computational Biology

Background:

• Epigenome-wide association studies (EWAS) in blood are crucial for understanding disease mechanisms.
• Existing DNA methylation reference libraries may not fully capture biological variability in blood samples.

Purpose of the Study:

• To investigate if augmenting existing DNA methylation reference libraries enhances the explanation of variation in blood-based EWAS.
• To compare the predictive performance of existing versus enhanced libraries in explaining methylation variability.

Main Methods:

• Utilized three publicly available 450K methylation datasets from whole-blood samples.
• Developed and compared enhanced DNA methylation libraries against existing ones.
• Modeled variability at each CpG site, adjusting for cell composition using both library types.

Main Results:

• An enhanced DNA methylation library explained significantly more variation across all three datasets compared to existing libraries (p < 10(-3)).
• The enhanced library reduced CpG-specific residual sums of squares error, indicating improved model fit.

Conclusions:

• Incorporating pathologically important immune cell subtypes into reference libraries is key for explaining variability in blood EWAS.
• Enhanced DNA methylation libraries offer a more complete explanation of variation in blood-based EWAS.