EWASex

EWASex predicts biological sex from DNA methylation data to enable accurate accounting of sex in epigenome-wide association studies.

Key Features:

• High accuracy: Uses DNA methylation at selected CpG sites on the X-chromosome that are stable under X-chromosome inactivation in females to enable precise sex estimation.
• Performance superiority: Demonstrated to outperform existing state-of-the-art methods across diverse EWAS datasets.
• Pre-trained weights: Includes pre-trained weights that permit sex prediction without access to RAW data.
• Tissue generalizability and label independence: Applicable to blood and other tissue types and functions without requiring prior sex labels.

Scientific Applications:

• Covariate adjustment in EWAS: Inferring sex to enable correct adjustment for sex as a variable and reduce bias in EWAS statistical analyses.
• Interpretation of epigenetic associations: Determining sex to support interpretation of DNA methylation differences associated with diseases or health traits.

Methodology:

Uses DNA methylation patterns at selected CpG sites on the X chromosome that are stable under X-chromosome inactivation and applies pre-trained weights for sex estimation.