Validating the Accuracy of a Rapid Phenotyping Paradigm Using Web-Based Parent Input

Background: Autism is known to be substantially inherited but specific molecular genetic abnormalities have only been identified for a small minority of all cases, indicating that there may be a large number of distinct genetic risk factors, each with relatively small explanatory power, either in the form of many rare strong effect variants, or common variants of modest effect. In order to delineate the genetic component of autism and identify specific genetic variants, sample sizes much larger than those currently under study will be needed.  Despite tremendous efforts, only 3-4000 European-descent autism families are currently available for genetic study even with 15-20 years of sample collection and phenotyping effort.

Objectives: The objective of this study is to demonstrate the accuracy of a rapid phenotyping paradigm consisting of brief parent- and teacher-report questionnaires, and medical record documentation of autism spectrum disorder (ASD) diagnosis, in order to minimize the costs and time of sample recruitment in the interest of more rapidly increasing the available sample size.

Methods: Families were invited to participate via the online Interactive Autism Network (IAN). The IAN registry, which grows at 300 individuals per month, currently has over 11,000 families registered with at least one autistic child: less than 5% have participated in genetic studies, but indicate willingness to participate. In order to assess the diagnostic specificity of the recruitment procedure, an email recruitment letter was sent to ASD-affected families in which the probands were: verbal (this is the case for 74% of the families in IAN); age 4-18 years; living near UCLA, Kennedy Krieger Institute or Washington University, St. Louis; had Social Communication Questionnaire (SCQ-lifetime version) scores greater than or equal to 12 at the time of entry.  These families were invited to one of the three clinics and assessed by ADI-r (Autism Diagnostic Interview-Revised), VABS (Vineland Adaptive Behavior Scale), an observational confirmation procedure (Autism Diagnostic Observation Schedule—ADOS—at 2 sites, expert-clinician observation at the third), followed by clinician best estimate diagnosis (BE).

Results: Interim analysis of the sample set was performed after 47 families were enrolled and assessed. 100% were ASD positive by BE and 100% were ASD positive by developmental history on the ADI-r.  91.5% were also positive by ADOS / clinical observation (this figure improved to 95% when excluding patients on the basis of sub clinical teacher-report Social Responsiveness Scale (SRS) score), suggesting that a few of the children with an established historical diagnosis may have substantially improved by the time of this study. There were no significant demographic or phenotypic differences between the enrolled probands and the larger population of verbal ASD subjects registered in IAN.

Conclusions: These data support the reliability of a rapid phenotyping paradigm for verbal subjects with ASD, and constitute a fundamental aspect of feasibility for large-scale efforts to advance—by an order of magnitude—the world collection of family-based biomaterials for autism genetic research.