Our data suggest a potentially important role for recessive mutations in autism. Though our pre-selection of 16 patients for whole exome sequencing, and our limited analysis of whole exome data from >400 cases in the ARRA Autism Sequencing Consortium, does not allow us to calculate the proportion of cases likely attributable to recessive as opposed to other causes (e.g. de novo, X-linked), our data do suggest that a systematic analysis of recessive causes of autism would be worthwhile. Homozygous null mutations appear to be exceedingly rare in autism, while homozygous missense changes were found in several candidate genes (Table 2), consistent with the possibility that some cases of ASD may reflect hypomorphic mutations in genes that have more severe phenotypes when completely disabled [11]. On the other hand, compound heterozygous recessive mutations could be more common in the outbred families represented by the AGRE.
