Autism spectrum disorders (ASDs) are incompletely understood neurodevelopmental disorders diagnosed solely on the basis of behavioral assessments of social, communicative, and repetitive symptoms.1 Although ASD is behaviorally distinctive and reliably identified by experienced clinicians, the disorder is clinically and genetically extremely heterogeneous. Psychologists, who began to define autism subgroups in the 1990s, found neither behavioral measures of core ASD symptoms nor cognitive measures reliably identified subgroups with similar outcomes or risk of recurrence in siblings.2 At the same time, geneticists were attempting with negligible success to find “autism genes” using molecular linkage and association analysis that had successfully identified the genes for cystic fibrosis and Huntington disease.3 Above all, their failures were attributed to an inability to assemble homogeneous ASD cohorts for analysis. This spurred the search for biomarkers, sometimes called endophenotypes, that might sort out the etiologic heterogeneity associated with ASD.4 The goal was to identify features that occur consistently in a portion of patients with ASD and are relatively discrete, quantifiable, and most importantly etiologically relevant. Physical dysmorphology appeared to fit the criteria.
Miles JH. Complex Autism Spectrum Disorders and Cutting-Edge Molecular Diagnostic Tests. JAMA. 2015;314(9):879-880. doi:10.1001/jama.2015.9577