|
Abstract:
|
Genome-wide methylation arrays quantify the methylation status of CpG dinucleotides sampled from across the human genome. Whole-genome approaches yield intensity values for millions of loci, including loci located not just within promoters, but also within repetitive sequence, gene bodies and intergenic sequence. The prioritization of loci for subsequent validation with a gold standard bisulfite-based analysis will depend heavily upon the observed information for each locus with respect to disease. However, ranking loci solely based on statistical significance treats all loci as equal, ignoring possible differences in the prior probability of a true association with disease and the expected magnitude of effect. We will illustrate one method for integrating prior information into a genome-wide methylation study and discuss both the biological and statistical implications of this approach.
|
