The analysis of DNA methylation patterns has become of great interest as methylome alterations have been found in many diseases. MeDIP (Methylated DNA ImmunoPrecipitation) immunoprecipitates genome-wide methylated sequences many of which are located in the repetitive sequences. Such sequences are difficult to align unambiguously after sequencing (MeDIP-Seq) leading to a large number of sequences that are currently not used for further analysis. We present an innovative method called MeDIP-dep-Seq which depletes a significant part of several classes of these highly repetitive sequences (up to 300-fold decrease), while unique sequences of interest are not affected. After sequencing on a second generation sequencer (GAIIx, Illumina) the alignment rate substantially enhanced increasing thus the amount of usable sequences. We have further developed a pipeline for the analysi of MeDIP-Seq datasets. Potential candidate regions identified in this genome-wide assay can then be validated by the use of selector probes that specifically capture genomic regions of interest. We introduced a bisulfite treatment in the selection protocol and developed a novel multiplex assay. 98 gene loci were enriched in 6 samples and were then sequenced in parallel on a bench sequencer (GS Junior, Roche). The combination of these technologies will permit the establishment of methylome maps and the identification of novel epigenetic biomarkers for cancer and complex diseases diagnostics and prognosis.
