Dennis Wall, associate professor of pediatrics at the Stanford University School of Medicine, is leading the “largest-ever collaborative, open-access repository of bioinformatic data on autism.” According to Wall, his goal is to figure out what causes autism and to find ways to develop therapy techniques for the condition.

The project is funded by a $9 million grant from the Hartwell Foundation. The goal of the database is to tackle the genetic component of autism. The online database will include visible observations of autism like social behavior and language as development as well as non-visible aspects such as data on genomes and genes.

The United States Centers for Disease Control and Prevention estimates that today one out of 68 people are born autistic in the United States – 30 percent higher than previous estimations in 2012. In addition, this increasingly prevalent condition still holds a high level of perplexity.

Wall explained that autism is multigenic, or controlled by multiple genes; however, genetics is not the only factor that leads to autism. In fact, researchers have estimated that only about 50 percent of autism behaviors are controlled by genes. Wall’s work focuses on studying the 50 percent of autism cases that people cannot yet explain.

“A large foundation of what is at the root of autism is likely to lie in the genomes,” Wall said. “Once we figure that out, we can parse out more complex aspects of autism that include environment.”

Michael Snyder, Stanford Ascherman Professor and Chair of Genetics and director of the Center of Genomics and Personalized Medicine, is also collaborating with Wall on the database. According to Snyder, uncovering the complexities of autism is “going to take a lot more than [looking at] genetics.”

Wall believes a significant variable in the unknown 50 percent that causes autism is related to environment. Environmental factors include elements such as exposure to pesticides, the age of an individual’s parents and the flora in the area where a person is living. Since these factors are harder to observe and connect with autism, most autism research studies genomes, Wall said. However, with over three million DNA base pairs in a whole genome sequence, the search for biomarkers indicating autism is not an easy task.

“We have had a difficult time discovering biomarkers, therapeutic intervention, over the last decades,” Wall said. “One of the reasons for this is the lack of numbers of data on a large number of individuals.”

Wall’s project has already shown success in detecting autism from young ages. Currently, the clinical intake process can only be marked by professionals. However, the amount of kids at risk of autism far exceeds the number of professionals who can diagnose autism.

“What you wind up with, just as a blatant math, is long, long waiting rooms,” Wall said. “Kids are waiting to get an assessment for a diagnosis.”

Currently, the average age for getting a diagnosis is four and a half, but by that time children have lost opportunities for therapeutic interventions that are more effective at a younger age. Wall and his team, however, have created a quicker and more effective way of detecting autism.

“Using machine learning and artificial development, we discovered an algorithm that can detect the presence and risk of autism in minutes using a mobile device and a short home video recorded by the parents of the child,” Wall said.

Through the home video, they look for play behaviors, eye contact, level vocabulary development and interactions with toys. From these observations, they are able to instantaneously detect autism – in under five minutes. Wall hopes to reduce the current standard of diagnosis from four and a half to two years old. The earlier autism is caught, the easier it is to treat, he explained.

Snyder acknowledged that addressing a complex issue like autism requires a level of creativity, an ability to observe autism in a unique way. He was impressed with the originality of Wall’s home videos.

“Autism is a complex disorder,” Snyder said. “No [single] person is going to solve it. We are coming at this problem from a different angle. Ultimately, we are trying to make sense of all the data, trying to turn the data into knowledge.”

Wall admits that it is unlikely to develop diagnosis markers and have clear targets for therapeutic interventions in the next three years. However, he is hopeful that this database can translate into learning of not only autism, but also other diseases.

“This database can be leveraged to understand the causes of other diseases,” Wall said. “Because human diseases are highly innervated, we can leverage the commonality in autism with a number of overlapping diseases.”

 

Contact Avni Prasad at avniprasad ‘at’ gmail.com.