Genome Wide Analysis of DNA Methylation

点击翻译按钮获取中文摘要

DESCRIPTION (provided by applicant): This proposal aims to identify functional, non-coding elements in eukaryotic genomes by surveying genome-wide DNA methylation. Cytosine DNA methylation controls developmental gene expression in mammals during genomic imprinting and X-chromosome inactivation, and also silences transposons and other repeated DNA sequences. Hypermethylation of specific tumor suppressor genes contributes to cancer, showing the relevance of this work to human health. Here we propose using whole-genome microarrays (WGAs) to find all methylated sequences in a eukaryotic genome; the "methylome". We will test our methods with the model plant Arabidopsis thaliana, demonstrating their usefulness for the human genome. Arabidopsis is ideal for studying DNA methylation, because it has facile genetics, a small genome, and orthologs of every human DNA methyltransferase. In contrast to the lethality of mouse DNA methyltransferase mutants, Arabidopsis can tolerate mutations that virtually eliminate methylation. Our Arabidopsis thaliana arrays tile the complete genome with 25-mer oligonucleotides, allowing us to precisely measure DNA and RNA hybridization. We will use several different reagents to identify methylated DNA sequences, each coupled with high-throughput analysis using WGAs: bisulfite treatment of genomic DNA, restriction digest with methylation-sensitive enzymes, anti-methylcytosine antibodies, and proteins that bind specifically to methylated DNA. By using complementary and independent methods, we hope to detect DNA methylation with up to single nucleotide resolution. Arabidopsis DNA methyltransferase mutants have well-characterized DNA methylation defects at several endogenous loci. We will use mutants in every major Arabidopsis DNA methyltransferase to test and verify DNA methylation detection methods. RNA silencing mutants will determine how much DNA methylation is guided by small intefering RNAs (siRNAs). To further define functional non-coding elements in the genome, we correlate changes in DNA methylation with transcription and with the presence of antisense gene transcripts, and other non-coding RNAs. Once we have identified all methylated loci, we will use novel sequence libraries to classify them as transposons, other repeats, or unique sequences. This analysis is likely to reveal functional non-coding elements that are invisible to other methods. Since methylation is critical in many mammalian gene regulation phenomena, the methods developed in this proposal will clearly move the ENCODE project toward its goal of identifying functional elements in the human genome.