NEW YORK (GenomeWeb) – Sites of differential DNA methylation and genetic associations for schizophrenia may overlap, according to a new study appearing in Genome Biology.
Researchers led by the University of Exeter Medical School's Jonathan Mill performed a multi-stage epigenome-wide association study that examined genome-wide patterns of DNA methylation in more than 1,700 people, both with and without schizophrenia. From this, they uncovered some 350 differentially methylated regions and positions in people with schizophrenia, about two dozen of which colocalized with genetic sites linked to schizophrenia through genome-wide association studies.
"By aligning the results from these two molecular approaches, we have generated a list of genes directly affected by schizophrenia genetic risk factors," first author Eilis Hannon from Exeter said in a statement.
In the first stage of their EWAS, the Exeter-led team examined the methylation status of a discovery cohort of 675 people, about half with and half without schizophrenia. After controlling for smoking status —smoking is more common among people with schizophrenia than the general population — and other confounding factors, the researchers uncovered 25 differentially methylated positions (DMPs) associated with schizophrenia at their most stringent threshold and 1,223DMPs at a more relaxed threshold.
To extend that collection, the team took two alternative approaches, using comb-p algorithm and sliding window approaches to link an additional 85 differentially methylated regions (DMRs) to schizophrenia. The researchers noted that the top DMR identified this way spanned three probes within the GYG1 gene, which has been found to be differentially expressed in prefrontal pyramidal neurons from people with schizophrenia.
Hannon and her colleagues then tested those differentially methylated positions and regions in a further 847 individuals and 96 monozygotic twin pairs. A meta-analysis of all three phases found that 22 DMPs from phase 1 had experiment-wide significance across all cohorts, and yielded a further 343 experiment-wide DMPs.
The locations of these DMPs also overlapped with some of the 105 autosomal genomic regions linked to schizophrenia through a recent GWAS conducted by the Psychiatric Genomics Consortium, the researchers reported. They found that 27 of the 76 GWAS regions they tested showed significant differences in DNA methylation. They also noted that these DMRs had combined p-values that exceeded the best DMP p-value, indicating that there might be more than one semi-independent differentially methylated spots in the regions.
These and other data support the notion that schizophrenia-associated DNA methylation differences overlap with genetic susceptibility loci, the researchers wrote.
"This study highlights the power of integrating different types of genomic data to better understand how disease-associated DNA sequence variation actually influences the way in which genes function," senior author Mill said in a statement. "Although our study focused on schizophrenia, were now applying this approach to other types of complex disease."