horizons of psychiatric genetics and epigenetics: where are we and where are we heading?

today multinational studies using genome-wide association scan (gwas) for >1000,000 polymorphisms on >100,000 cases with major psychiatric diseases versus controls, combined with next-generation sequencing have found ~100 genetic polymorphisms associated with schizophrenia (scz), bipolar disorder (bd), autism, attention deficit and hyperactivity disorder (adhd), etc. however, the effect size of each genetic mutation has been generally low (<1%), and altogether could portray a tiny fraction of these mental diseases. furthermore, none of these polymorphisms was specific to disease phenotypes indicating that they are simply genetic risk factors rather than causal mutations. the lack of identification of the major gene(s) in huge genetic studies increased the tendency for reexamining the roles of environmental factors in psychiatric and other complex diseases. however, this time at cellular/molecular levels mediated by epigenetic mechanisms that are heritable, but reversible while interacting with the environment. now, gene-specific or whole-genome epigenetic analyses have introduced hundreds of aberrant epigenetic marks in the blood or brain of individuals with psychiatric diseases that include aberrations in dna methylation, histone modifications and microrna expression. interestingly, most of the current psychiatric drugs such as valproate, lithium, antidepressants, antipsychotics and even electroconvulsive therapy (ect) modulate epigenetic codes. the existing data indicate that, the impacts of environment/nurture, including the uterine milieu and early-life events might be more significant than genetic/nature in most psychiatric diseases. the lack of significant results in large-scale genetic studies led to revise the bolded roles of genetics and now we are at the turning point of genomics for reconsidering environmental factors that through epigenetic mechanisms may impact the brain development/functions causing disease phenotypes.