De novo rates and selection of schizophrenia-associated copy number variants

Abstract

Background: At least 10 large and rare recurrent DNA copy number variants (CNVs) have been identified as risk factors for schizophrenia and other neurodevelopmental disorders. Because such conditions are associated with reduced fecundity, these pathogenic CNVs should be filtered out from the population by selection and must be replenished by de novo events.

Methods: To estimate the mutation rate (μ) for these CNVs and the selection pressure (s) against them, we first conducted a literature review on the rate of each of these CNVs in the population and the rate of their de novo occurrence. In each generation, the number of CNVs lost because of reduced fertility must be replenished by the same number of de novo CNVs. Therefore, the observed ratio of de novo versus all (inherited + de novo) CNVs approximates the selection coefficient (s) of that CNV. The mutation rate approximates to μ = s × q, where q is the frequency of the CNV in the population.

Results: High selection pressure operates at all these loci (s = .12 - .88), suggesting that following de novo occurrence, each of these CNVs persists in the population in only a few generations. The mutation rate for each CNV is high, affecting between 1:3500 and 1:30,000 individuals. The rarest CNVs have the highest selection coefficients.

Conclusions: The CNVs that increase risk to develop schizophrenia are caused by recent de novo mutations and are under strong selection pressure. They persist in the population because of high mutation rates.