New functional analysis can predict and rank autism genes
A new computational analysis can predict whether a gene is implicated in autism or intellectual disability with up to 98 percent accuracy, according to a study published 15 May in the American Journal of Medical Genetics1.
In the new study, researchers created three computational classifiers using data for 114 candidate genes for autism and 223 for intellectual disability to determine the probability that a gene is linked to either of the disorders. Two of these classifiers are based on protein-protein interactions, and the third uses functional information, such as which pathways the gene is involved in, and what effect deleting the gene has on mice.
The functional classification is the most accurate, distinguishing autism and intellectual disability genes from a set of control brain-expressed genes with 80 to 98 percent accuracy, the study found.
The system is also better able to discriminate between autism genes and control genes expressed in the brain than it is between intellectual disability genes and control genes. This suggests that genes implicated in intellectual disability influence a broader range of neurological pathways.
The analysis also shows that intellectual disability and autism genes are more closely linked to each other than would be expected based on chance alone, suggesting that they belong to related pathways.
1: Kou Y. et al. Am. J. Med. Genet. C. Semin. Med. Genet. 160, 130-142 (2012) PubMed