In this article, we review studies detailing the correspondence between peripheral blood and brain tissue across various domains of high-throughput -omic analysis in order to provide a context for evaluating blood-based biomarker studies. Specifically, we reviewed seven studies comparing patterns of DNA methylation (i.e., an aspect of the epigenome), eight articles comparing patterns of gene expression (i.e., the transcriptome), and three articles comparing patterns of protein expression (i.e., the proteome). Our review of the epigenomic literature suggests that CpG-island methylation levels are generally highly correlated (r ¼ 0.90) between blood and brain. Our review of transcriptomic studies suggests that between 35% and 80% of known transcripts are present in both brain and blood tissue samples; estimates of cross-tissue correlation in expression levels were found to range from 0.25 to 0.64, with stronger correlations observed among particular subsets of genes. Relative to the epigenome and transcriptome, the proteome has not been as fully compared between brain and blood samples, highlighting an important area for future work as whole-proteome profiling methods mature. Beyond reviewing the relevant studies, we discuss some of the assumptions, methodological issues, and gaps in knowledge that should be addressed in order to better understand how the multiple ''-omes'' of the brain are reflected in the peripheral blood. A better understanding of these relationships is a critical precursor to the validation of biomarkers for brain disorders. Ó 2013 Wiley Periodicals, Inc.Key words: blood; brain; epigenome; gene expression; genome; methylation; neuropsychiatry; proteome; transcriptome
INTRODUCTIONTechnological advances in molecular biology over the last two decades have fundamentally altered our approach to studying brain disorders. The ''throughput'' of many molecular-profiling techniques, which had heretofore been a major bottleneck to discovery, has increased exponentially. This, in turn, has initiated a notable trend in the field away from hypothesis-testing of candidate genes, transcripts, and proteins, and toward hypothesis-generation through the simultaneous evaluation of all members of a particular molecular species (i.e., an ''-ome''). As a consequence, the number of -omes entering the scientific vernacular has also increased rapidly to include the genome (all DNA sequence variations), epigenome (all chemical modifications to the DNA and histone proteins), transcriptome (all expressed RNA transcripts), and proteome (all expressed proteins), among others.Given the demonstrable heritability of many neuropsychiatric disorders [Glatt et al., 2010], much research has been dedicated to identifying the genomic variations underlying their susceptibility. These genomic studies have been predicated on the fundamental and largely true assumption that the genome is equivalent in sequence and structure in all cells and tissues of the same organism. This tenet allows for the standard practice of examining DNA sequen...