Blood-based multivariate methylation risk score for cognitive impairment and dementia

Area of Science:

• Epigenetics and Neurodegenerative Diseases
• Biomarker Discovery
• Computational Biology

Background:

• DNA methylation is linked to dementia pathophysiology and may mediate lifestyle/environmental impacts.
• Blood-derived DNA methylation offers a promising avenue for early dementia risk detection.
• Understanding epigenetic modifications is crucial for unraveling dementia's complex etiology.

Purpose of the Study:

• To investigate the utility of blood DNA methylation as a surrogate for dementia risk factors.
• To develop and validate a predictive model for cognitive impairment using methylation data.
• To assess the potential of machine learning in conjunction with omics data for dementia risk assessment.

Main Methods:

• Genome-wide DNA methylation data from whole blood was analyzed.
• Machine learning techniques were employed to build a risk prediction model.
• The model used DNA methylation as a surrogate for 14 modifiable and non-modifiable dementia risk factors.

Main Results:

• A multivariate methylation risk score (MMRS) was developed to identify mild cognitive impairment (MCI) cross-sectionally.
• The MMRS independently predicted the prospective development of cognitive impairments in Alzheimer's and Parkinson's disease cohorts.
• The score demonstrated significant predictive power for cognitive decline (e.g., RAVLT-Learning and MCI/dementia).

Conclusions:

• Blood-derived DNA methylation data holds significant potential for dementia risk assessment.
• The developed MMRS can predict cognitive impairment at a population level.
• This approach highlights the synergy of omics data and machine learning in advancing dementia prediction.