Unifying DNA methylation-based in silico cell-type deconvolution with deconvMe.

Area of Science:

• Epigenetics and Computational Biology
• Single-cell analysis

Background:

• Cell-type deconvolution is crucial for analyzing complex biological samples.
• Existing methods for DNA methylation deconvolution are less accessible compared to gene expression data.
• Understanding cellular composition is vital in various biological contexts.

Purpose of the Study:

• To introduce deconvMe, an R package simplifying DNA methylation-based cell-type deconvolution.
• To provide accessible tools for researchers working with DNA methylation data.
• To compare deconvolution estimates from DNA methylation data with gene expression and ground truth data.

Main Methods:

• Development of the deconvMe R package.
• Application of deconvMe to DNA methylation data from blood samples.
• Comparison of deconvMe estimates against gene expression deconvolution and experimental ground truth data.

Main Results:

• The deconvMe package offers simplified access to DNA methylation deconvolution methods.
• Comparisons reveal the performance of methylation-based deconvolution for blood cell types.
• The study provides a unique matched dataset for evaluating deconvolution methods.

Conclusions:

• deconvMe enhances accessibility to DNA methylation deconvolution for biological research.
• The package facilitates robust cell-type estimation from epigenetic data.
• This work supports the use of DNA methylation for deconvolution, particularly in blood studies.