Using epigenetic data to estimate immune composition in admixed samples.

ABSTRACT

Variation in cell-type decomposition can be of interest to changes in biology by themselves, or a confounder in comparing admixed samples with different phenotypes. Here, I present and review various algorithms that may be used for purposes of deconvolving epigenomic data from admixed samples to yield estimates that may be factored into downstream analyses. More specifically, I present a protocol based on an adaptation of the CIBERSORT algorithm, previously developed for gene-expression, to methylation data.