A multitude of tools for pathway analyses are available. These differ GW-572016 order importantly, including the null hypothesis tested and whether they provide self-contained tests or competitive tests. Self-contained testing provides evidence for association of gene-sets (or genes) with a trait, regardless of association of other gene-sets (or genes). The null hypothesis in self-contained testing is that there

is no association with the trait in any gene-set (or gene). Self-contained testing thus does not take into account a possible polygenic nature of a trait and therefore might produce biased results, as the null hypothesis strictly is not true for such traits. In addition, the self-contained test is sensitive to spurious association, which thus needs

to be taken into account for example in the data cleaning steps. Competitive testing evaluates association of a gene-set with a trait while in relation to associations of other gene-sets. The null hypothesis for a competitive test is that the selected gene-set is not more strongly associated to the trait than any other (matched) set of genes, thus correcting for an overall, polygenic effect on association. The competitive test is not sensitive to spurious association due to population stratification. Competitive tests tend to be more conservative than self-contained tests, but are generally preferred. Wang et al. [25•] provide an excellent overview of these methods. In this review, we summarize the evidence for the involvement of biological pathways for multiple psychiatric disorders (Table 1). We restrict SB203580 cost ourselves to studies based on whole genome approaches: GWAS, copy number variant (CNV), or exome sequencing data. We exclude studies based solely on gene expression data or pharmacological interventions. We cover gene-sets that reflect protein–protein interaction (PPI) networks,

expert curated sets of functionally isothipendyl related genes, sets of co-expressed genes, and traditional pathways (e.g., genes involved in the synthesis, storage, release, binding, and degradation of a neurotransmitter). We focus on five psychiatric disorders — SCZ, ASD, BD, MDD, and ADHD. SCZ has been the focus of many published pathway analyses. Given the critical importance of sample size, we focus on the largest and most recent studies. The largest GWAS for SCZ was from the Psychiatric Genomics Consortium (PGC) [17••] and included previously reported pathways. This paper reports evidence for genes whose mRNAs bind to FMRP [12••], genes containing a predicted target site of miR-137 [26], and neuronal calcium signaling. A role for immune function was suggested by the prominence of the human leukocyte antigen (HLA) genes and enrichment of GWAS associations in epigenetic enhancer regions. Although not studied in the PGC paper, prior results consistently implicated synaptic 27 and 28], astrocyte 29 and 30••] and oligodendrocyte 29, 30•• and 31] biological processes.