28 - IMAGE-CpG: DEVELOPMENT OF A WEB-BASED SEARCH TOOL FOR GENOME-WIDE DNA METHYLATION CORRELATION BETWEEN LIVE HUMAN BRAIN AND PERIPHERAL TISSUES WITHIN INDIVIDUALS

Abstract

Background Differential DNA methylation (DNAm) in the brain is associated with a number of psychiatric diseases, but access to brain tissues is essentially limited to post-mortem samples. The use of surrogate tissues has become common in identifying methylation changes associated with psychiatric disease. We determined the extent to which these peripheral tissues can be used as surrogates for assessing DNAm in the brain. Because of its potential value for researchers, a web-based tool has been developed to facilitate the evaluation of the peripheral tissue to brain correlation for CpGs across the genome. Methods DNA from blood, saliva, buccal, and live brain tissue samples from 17 treatment-refractory epilepsy patients undergoing brain resection were collected. Genome-wide DNAm was assessed using the Infinium HumanMethylation450K BeadChip array on 10 subjects, initially, for brain, blood, and saliva samples. The Infinium EPIC array was used to analyze 13 subjects with brain, blood, saliva, and buccal samples, with six subjects overlapping between the two datasets. Data were processed and analyzed with the R packages Minfi and RnBeads. Tissue relatedness was analyzed with Pearson correlation. Results From the EPIC methylation data, the saliva-brain correlation (r2=0.79) was higher than that for blood-brain (r2=0.74) and buccal-brain (r2=0.71). Genomic region comparisons showed that correlations within promoter (saliva r2=0.88, blood r2=0.84, buccal r2=0.83) and genic regions (saliva r2=0.82, blood r2=0.77, buccal r2=75) were higher than within intergenic regions (saliva r2=0.67, blood r2=0.60, buccal r2=0.55). To allow for detailed and gene-specific use of this data, a website tool was created that displays both Infinium 450 K and EPIC data. The user chooses between platforms, and then inputs either an individual CpG, gene name, or chromosomal region for which to receive information. The tool then generates as output a figure comparing the DNAm levels within the tissues, along with the degree of correlation across that region. A table is also provided that includes the DNAm data across the samples tested, mean DNAm for each tissue, and the level of significance of the correlation at each CpG included in the input. Discussion Genome-wide analysis of DNAm from saliva, buccal, and blood tissues all revealed a high degree of correlation with live brain DNAm within individuals. Although overall saliva was most highly correlated, across some genes or CpGs buccal tissue or blood were more highly correlated. This indicates that to find the most useful surrogate for assessing DNAm in the brain, the level of correlation at a specific CpG, gene, or region must be considered. To assist in that objective, a website has been developed that allows researchers to interrogate DNAm levels and degree of correlation in user-defined locations within the genome.