Large sums of research funding are currently spent on trying to understand the biological and neurological profiles associated with autism, whereas there has been a relative paucity of funding for studies into service, intervention and societal issues. Heterogeneity of etiology may be one reason why progress has been slow; many people now talk of ‘the autisms’ to reflect the belief that different individuals have different biological paths to autism. Autism is due to a mixture of common inherited genetic variation across many genes each of small effect and rare mutations of large impact. The field of genetics moves very fast, and specific information about candidate genes would soon date and be unhelpful. Although the genetic contribution to autism is large, even identical twins do not show 100% concordance, and gene-environment interactions may be hidden in traditional estimates of heritability. Such biomarkers might be used to identify autism pre-natally or in infancy, or to identify sub-types of autism.