An Efficient Approach to Screening Epigenome-Wide Data