Skip to main content
Premium Trial:

Request an Annual Quote

SPARK Autism Research Initiative Obtains First Results, Explores New Recruitment Strategies


This article has been updated to correct that the arbaclofen clinical trial is sponsored by Clinical Research Associates, an affiliate of the Simons Foundation, and to include additonal information.

NEW YORK – A little more than three years after its launch, the SPARK research initiative, which is investigating genetic and non-genetic causes of autism, has grown to more than 220,000 participants from 75,000 families living in the US.

Run by the Simons Foundation Autism Research Initiative and supported by a multi-million dollar budget, SPARK has already exceeded its original recruitment goal of 50,000 families and continues to grow.

"To my knowledge, it's the largest autism study in the United States," and possibly in the world, said Wendy Chung, the study's principal investigator and a clinical geneticist and professor of pediatrics at Columbia University.

SPARK, which stands for Simons Foundation Powering Autism Research for Knowledge, has already generated a number of genotyping, exome sequencing, and whole-genome sequencing datasets, which it is making available to the autism research community in a de-identified manner. Whole-genome sequencing for the project is provided by the New York Genome Center and exome sequencing by the Regeneron Genetics Center.

In August, the project published a pilot study with exome sequencing results from more than 450 SPARK families in NPJ Genomic Medicine. One goal of that project was to show that generating genomic data from saliva samples worked as well as from blood samples, "and the answer was, it worked beautifully," Chung said. Another goal was to demonstrate that genetic findings from SPARK, which lets autism families enroll directly, are broadly similar to studies that verified an autism diagnosis independently using standardized measures, like the Simon Simplex Collection. "We didn't do any of that [for SPARK]. Essentially, we have to trust people that they are correctly identifying themselves as 'autism,'" Chung said.

The mutation spectrum and mutation frequencies in the pilot cohort were indeed similar to previous findings from the Simon Simplex Collection. In about 10 percent of families, the SPARK team identified variants in known autism genes, and in another 3 percent variants possibly associated with autism. "You could say, 'ho-hum, you mostly found the same things that had been seen before,' but that was actually the point, to make sure we were seeing the same things that had been seen before," Chung said.

The pilot also discovered some new genes that have since been confirmed as autism and neurodevelopmental genes, Chung said. In addition, it identified new autism candidate genes that are still under investigation.

In terms of the biological pathways those new genes belong to, "I don't think there was anything that was truly unexpected," she said. They include genes encoding transcription factors, RNA binding proteins, and proteins involved in splicing and signaling pathways. In addition, the researchers found genes encoding cell adhesion proteins and proteins with a role in cell migration, as well as genes whose products are involved in synaptic function and neurotransmitter production. "They are all genes that, to me, make sense, and to a certain extent, we've seen genes within those pathways or networks before," Chung said. "But this [list] is now expanding, and I hope will allow us to see some things downstream."

Developing new autism treatments that are based on genetic mutations will depend on the nature of the specific alteration, for example, whether it leads to a loss of function that can be restored or a gain of function that could be knocked down. In addition, researchers will have to prioritize potential targets by focusing on mutations that have reversible effects, she said. "If this is a done deal by the time you're born, there is a different window of opportunity for treatment than if this is something where you have a window up until 10 years of age," Chung said. Also, successful therapies may not work on a gene-by-gene basis but may target common pathways downstream.

Following the pilot study, a whole-genome sequencing dataset for 600 family quads – parents and two siblings – is scheduled to be released in the near future, and a second WGS dataset from more than 1,000 autism families will come out later in 2020, Chung said.

In addition, more than 6,500 SPARK families have undergone exome sequencing and genotyping, and another dataset from about 5,000 families will be released in early 2020.

While it is unclear exactly how much overlap in enrollment exists between SPARK and other large-scale autism studies, such as the Autism Sequencing Consortium or the Simon Simplex Collection, SPARK has been able to address this for meta-analyses using genetic data. "There has been some overlap and we've been able to flag those cases, and for the genomics community, we've actually suppressed those," so they would not unintentionally be double-counted, Chung explained. This is important, she said, because "increasingly, the autism genomics community is grabbing large datasets and pooling them all together, and that's where the power is going to come from."

What is unique about SPARK is that researchers can recontact participants and that a high level of genomic detail is available for them. In many ways, this makes SPARK similar to the National Institutes of Health's All of Us research program, Chung said, which doesn't focus on a specific condition or disease. Also, autism researchers can use SPARK as a resource to recruit participants for IRB-approved studies, and more than 40 studies have already made use of this free research match service. "I would not say it's been maximally utilized yet because I think people are still finding out about it but it's been very effective for the researchers who have used it so far," Chung said.

Response rates to invitations for research studies vary, she said, ranging from more than 50 percent for online surveys to less than 10 percent for clinical trials of autism therapies that involve in-person visits.

SPARK also returns genetic findings related to autism to families, and so far, more than 150 families have received such a finding, a number that keeps growing. SPARK has a "high bar" for returning genetic results, focusing on highly penetrant genes with a major impact on the patient, Chung said, and the diagnostic rate has been about 8 to 10 percent. More than 50 percent of the results are de novo mutations, about half of them copy number variants and the other half single nucleotide variants, while the remaining ones are inherited mutations.

"There are a lot of things that are on our radar that we haven't yet started returning because we want to be able to get more confidence statistically, as well as more understanding of the conditions, so we've got useful information to give back to the families," Chung said.

Genetic results are communicated to families with the assistance of genetic counselors, she added, and surveys of participants have shown a high degree of satisfaction with and understanding of the results.

Having a molecular diagnosis can help families in several ways. "The most mundane is, it gives the family some answers," Chung said, and it can help them make reproductive decisions if the condition is inherited.

Also, the variant can point to other, unrecognized medical problems, such as epilepsy. Based on one patient's genetic result, for example, he had an EEG done, and "they saw there were seizures that were going on that were not recognized that were part of [the patient's] academic difficulties and some of the behavioral difficulties," Chung said. "So, that's been immediately medically helpful in some cases."

Variants in a small number of genes are also associated with an increased cancer risk. One gene the researchers have frequently seen mutations in is PTEN, for example, which increases breast cancer risk. "We haven't diagnosed any cancers, thankfully, yet," Chung said, "but it is changing the screening that the families are doing."

Yet other genetic variants point to a neurodegenerative course of the patient's condition. "Given the age distribution, most of the [patients] in SPARK are still children, and some of the decline isn't seen until adulthood, and we haven't seen that yet," Chung said. "But it does tell us about prognosis and gives those families a certain sense of urgency in terms of trying to move forward with treatments."

For patients with one particular variant – the 16p11.2 microdeletion, which occurs in about 1 in 200 individuals with autism – Clinical Research Associates, an affiliate of the Simons Foundation, is starting a clinical trial for a medication called arbaclofen, in collaboration with three clinical sites around the country. In mouse models of this deletion, the drug has shown to be helpful for memory, social behavior, and motor impairments, which are also seen in people with the deletion. Also, the medication has a good safety profile and has been generally well-tolerated in previous studies of patients with autism or fragile X syndrome, Chung said.

While there are other genetics-based clinical trials of drugs for neurodevelopmental conditions, such as Angelman syndrome or Rett syndrome, this is CRA's first clinical trial for autism patients with a specific genetic change, she said.

Recruitment on the subway and on buses

SPARK is currently open to individuals with a diagnosis of autism spectrum disorder and their families living in the United States, who can join through 31 clinical sites around the country. 

The majority of autism cases in SPARK are school-aged children, with a mean age of nine years. But the study also includes one of the largest cohorts of adults with autism, Chung said, which keeps growing when patients who enrolled as children turn 18.

Over the last year, the project has focused on increasing recruitment from underrepresented minorities in order to be more representative of the US population, Chung said, and four of the recruitment sites have been conducting pilot projects on how to best engage underserved communities, including Latinos and African Americans.

In general, the 31 sites recruit patients from their own hospitals or clinics, as well as through community events for families affected by autism. In addition, SPARK has explored reaching out to families through the Internet and social media, including Facebook, YouTube, search engines, and banner ads. It has also tried traditional advertising, including on the New York City subway, on buses in Washington, DC, and signs on trains in Chicago. "We try a new recruitment strategy, and the good thing is, we can see numbers coming in instantaneously and we are able to pivot," Chung said. "For me personally, this has been a new opportunity to do that in a very data-driven way, and I think that's been very helpful."

One type of outreach that has been very effective, she said, is videos in which individuals, including families participating in SPARK, tell their stories and talk about their journeys. In particular, stories about autism families and researchers produced in collaboration with local news outlets have been very successful for enrollment. "You can see when a story runs and then a big spike happens in terms of recruitment right afterwards," Chung said.

The effectiveness of the transport ad campaigns are a little more difficult to measure, she said, as are some Internet ads that don't use tagged URLs. The most cost-effective way of recruitment has been Facebook ads, which also allow costs per referral to be capped at a certain amount.

While SPARK has surpassed its original recruitment goal of 50,000 families, it has been tricky to recruit family trios. "That's proven more challenging in the sense that we very often get the individual with autism, oftentimes one of the parents. But sometimes we have trouble getting the father in the family," Chung said. 

Part of the reason is that having a child with autism places a lot of stress on a family, leading to higher rates of separation and divorce. But even when fathers are present, they often have other priorities than enrolling in SPARK. "Oftentimes, there is one parent who is the driver and more often it's the mom who is the driver and the dad has other things going on," she said.

One general challenge with SPARK has been that autism is such a heterogeneous condition, affecting high-functioning individuals who live independently but have certain social challenges but also individuals who are non-verbal, have seizures, and show self-injurious behaviors. "It's hard sometimes to be able to message to the whole diversity of the community with one single voice and one single message, so we had to be quite careful about that and conscious that one message might be interpreted differently by different people in the community," Chung said.