UC San Diego

DNA methylation at CpG dinucleotides in mammalian cells is recognized as an epigenetic mechanism that plays a major role in mammalian development via gene expression regulation. Techniques in DNA methylation profiling have been advancing in the past decades. I have developed the second-generation of bisulfite padlock probe (BSPP) method, which does not require multiple steps of standard library preparation. This method is high-throughput and more scalable for quantification of DNA methylation at single-base resolution. The library-free method greatly reduces sample-preparation time and cost and is also compatible with automation. These developments have fulfilled the key requirements of a DNA methylation assay, including cost effectiveness, minimum sample input requirements, accuracy, and throughput. I have performed this technique to compare with other assays performed by different research groups for locus-specific DNA methylation analysis on the same samples set. BSPP assay showed a high correlation with other assays that have highest accuracy and is at the top with other assays based on the throughput. Genetic variants have an impact on local DNA methylation patterns by influencing methyltransferase recognition sequences or altering the DNA binding affinity of cis-regulatory proteins. To study this interaction, I have characterized CpG methylation state of 96 individuals from 22 nuclear pedigrees consisting of 52 parent-child trios using BSPP. I used the DMR330k probe set to quantify DNA methylation level at a set of 411,800 CpGs. Next, I have employed three independent approaches, including mid-parent offspring (MPO), methylation quantitative trait loci (mQTL), and allele-specific methylation (ASM) analysis, to investigate the influence of genetic polymorphisms on DNA methylation variation. MPO analysis identified 10,593 heritable CpG sites, among which 70.1% were SNPs that present in CpG sites. With mQTL analysis, 49.9% of heritable CpG sites were identified where regulation occurred in a distal cis-regulatory manner while ASM analysis was only able to identify 5% of heritable CpGs. This finding suggested that mQTL analysis do not identify all the cis-regulartory SNPs associated with heritable CpG methylation, and ASM analysis has even less power. I have extensively proved that in addition to regulating the mean of DNA methylation, genetic polymorphisms are also associated with the variability of DNA methylation levels. I have identified hundreds of CpG clusters in human genome for which the degree of DNA methylation variability was associated with genetic polymorphisms. This finding supported the previous studies showing that genetic variants have the influence on phenotypic plasticity such as gene expression or DNA methylation.