The dichotomy presented here between common SNPs of small effect, and rare but penetrant large CNVs, is an over-simplification. Exome sequencing (in which the expressed regions of each gene are sequenced in their entirety) is revealing a spectrum of schizophrenia-associated genetic variants (Gilman et al., 2012; Gulsuner et al., 2013), as there is with other psychiatric disorders (Visscher et al., 2012). In a recent large study, Purcell et al. (2014) show that schizophrenia is also associated with rare single nucleotide coding variants and by small insertions or deletions affecting a few nucleotides (‘indels’). This form of genetic variation was previously largely invisible. Purcell et al. (2014) estimate that such mutations account for a broadly comparable proportion of schizophrenia risk as do CNVs, with both contributing roughly one-tenth of the heritability attributable to common SNPs. Fromer et al. (2014) use family trios to show that these mutations are commonly de novo, and associated with more neurodevelopmental and cognitive impairment than cases without such mutations. It is important to note that these rare variant studies do not implicate conclusively any specific gene,
