The Wall Lab at Stanford University


The Big Picture Project: a family study of autism


Although Autistic Spectrum Disorder (ASD) is a strongly genetic condition, certain environmental causes and metabolic disturbances may contribute to clinical features. Leveraging social networks, our goal is to conduct a completely crowdsourced clinical trial that will enroll thousands of internet-active families with autism quickly, we will examine the role of the microbiome and its effects on host metabolism, immune and mitochondrial function, in regard to the patient genotype, use uniform phenotypic classification and determine the geographic distribution of the patients to perform a comprehensive study of autism.
This project has three components:

  • Microbiome
  • Video Based Classifier and Caregiver Directed Survey
  • GapMap

Specific Aims

Autism Microbiome

The existence of a link between the gut microbiome and ASD is well established in mice, yet its impact in humans remains poorly understood. Although previous studies have shown that individuals with autism had gut microbiomes following specific patterns, they have suffered several limitations lacking appropriately matched controls and not accounting for the fact that ASD is a continuous phenotype. In order to overcome these limitations, we have initiated a plan to sequence 16S amplicon and analyze the gut microbiome of pairs of siblings, each pair containing a child with autism aged between 2-7 years and his/her unaffected sibling within 2 years of age. We recruited nearly one hundred internet-active families, and confirmed participant’s phenotype by using two published machine-learning classifiers based on a brief parent questionnaire and an analysis of a home video. Our preliminary analyses on 50 samples show that the gut microbiome community preferably clusters between siblings of each family. A Between Group Analysis was computed using 3 groups (higher functioning participants, severe autism symptoms and the unaffected sibling), and allowed us to extract bacterial taxa specific to each group. We were also able to build a classifier using logistic regression and extract a group of 30 bacterial genera of importance to distinguish participant with ASD and their unaffected siblings with an accuracy of 76%. We are currently integrating the remaining samples to complete this mapping of the microbial diversity across ASD. We believe this crowdsource-driven approach will result in dramatic clinical advances. Families interested in participating are encouraged to visit the Autism Microbiome web site for additional information on how to get involved.

Video Based Classifier and Caregiver Directed Survey

We seek to compare the genotypic data from the Microbiome to behavioral, phenotypic data that will be captured from participants from this component.

GapMap: prevalence and geographic distribution of autism through an online and web accessible tool

We have developed a prototype tool called GapMap, intended to source a massive and growing community of families with autism.  GapMap collects locational, diagnostic, and resource-use information from individuals and/or families with autism in order to compute accurate prevalence rates and better understand autism resource epidemiology. The application works in three main ways:

(1) as an all inclusive and searchable resource database for the autism community;

(2) to track, report, and understand where families with autism live in comparison to where there are available autism-focused resources; and

(3) to benefit the research community by helping studies target and reach potential participants.