Copy number variation and chromosome rearrangements may be among the causes of autism.

Autism and autistic spectrum disorders are known to be highly heritable but the actual genes involved have been hard to pin down. In large part, this is because they are highly heterogeneous conditions - not one disorder but a multitude of distinct disorders sharing similar symptoms. Genome-wide association studies have identified some candidate genes - and analysis of chromosomal abnormalities is offering an alternative approach.

Autism-like behaviours are seen in many genomic disorders. If the chromosomal abnormality has been identified, then the genes affected are probably involved in the development of autism. With a translocation, a single gene may be affected; more usually a deletion will eliminate several genes. Researchers can look at the deleted genes and see which might be responsible for the symptoms of autism (e.g. if they are active in the brain).

Alternatively, array analysis may reveal copy number variation in affected individuals. Again, researchers can then narrow their hunt down to genes in the altered regions.

Through these approaches, many genes that may contribute to autism or autism-like conditions are being identified. It remains a significant challenge to confirm that they are important and to work out what they are doing. Crucially, it may be possible to piece together how a gene abnormality might affect behaviour. Several genes, for example, code for proteins known to be present at the synapse of nerve cells, so could be affecting how nerve signals are transmitted in the brain.

Given that the brain and behaviour are so complex, it is highly likely that many different genes are involved in autism, acting individually or in combination, and interacting with environmental factors. Even the most common genetic changes will probably account for only a tiny fraction of cases.