Credit: Getty

A clutch of recent papers have harnessed exome sequencing to study the contribution of single-nucleotide variants to the risk of autism spectrum disorder (ASD). Four studies have looked for de novo mutations, and a fifth looked at rare recessive mutations; together they highlight the extreme genetic heterogeneity of ASD and point towards biological pathways for further study.

A role for de novo mutations in ASD has been suggested by previous studies of copy number variation and smaller-scale exome sequencing studies. In the recent studies, Iossifov et al. sequenced 343 family 'quads' (the parents of a single child on the autism spectrum and its unaffected sibling), and Sanders et al. also included 200 quads in the 238 families they sequenced. O'Roak et al. selected 189 trios (a child with ASD and its parents), and Neale et al. also sequenced 175 trios. Although the details differ among the four studies, the findings reveal several shared patterns and interesting leads.

six genes were found to have a gene-disrupting de novo mutation in more than one individual with ASD

First, this wealth of sequencing data from parents and offspring enables estimation of mutation rates. The estimates for de novo point mutations are ∼2 × 10−8 per base per generation, which is slightly higher than in previous studies. Although the studies do not find elevated mutation rates in the probands compared with their unaffected siblings, evidence is presented of a shift in the mutation spectrum towards mutations predicted to disrupt protein function in the probands. In addition, the studies show that more de novo mutations have a paternal origin than have a maternal origin and that there is an increase in the number of mutations with paternal age.

Each of the studies identified de novo mutations that are predicted to disrupt gene function in some of the affected children: the total haul is 127 gene-disrupting mutations. Although only a proportion of these would be expected to be causal — Iossifov et al. estimate 65 of them — it is particularly noteworthy that six genes were found to have a gene-disrupting de novo mutation in more than one individual with ASD, and many others genes are 'hit' by missense mutations more than once. These will be important leads for follow-up work.

In terms of the biology of ASD, among the genes with de novo mutations are many that have been previously implicated in other neurodevelopmental disorders. Some analyses of the pathways in which the genes with de novo events lie also point to areas for further study. For example, O'Roak et al. identified a network linked to β-catenin signalling and chromatin remodelling, and Iossifov et al. highlighted many genes that are associated with the fragile X protein.

De novo mutations are just one piece in the ASD puzzle, and the work of Chahrour et al. describes an approach to access another difficult piece: recessive mutations. These authors sequenced the exomes of 16 probands from an outbred population who had abundant homozygosity in their genomes, which may reflect shared ancestry. They found four candidate autism genes with homozygous missense mutations.

Overall, these studies confirm estimates of several hundred genes being associated with ASD, and future work will need to bring together knowledge of both de novo mutations and inherited genetic variation.