Submission deadline: June 27, 2025
Earliest Possible Start Date: September 1, 2025
Funds available: $50,000,000 in FY25 with future funding subject to congressional appropriations and programmatic needs.
Health and Human Services Secretary Robert F. Kennedy Jr.
The purpose of this Research Opportunity Announcement (ROA) is to invite applications from eligible organizations to support the Autism Data Science Initiative (ADSI). ADSI will bring together diverse data resources and community members with lived experience to explore novel contributors and/or to characterize the collective contributions of numerous factors to the causation of autism spectrum disorder (ASD), hereafter referred to as autism, and their potential role in increasing the prevalence of autism. ADSI will also seek to identify how existing treatments/interventions are used and better understand their outcomes to inform the design of future clinical studies. This initiative will achieve these goals through four strategic aims: 1) to create new integrated data resources, by applying innovative approaches across existing data from research studies or other valid sources,1 with rigorous privacy protections, for use by the autism research community; 2) to identify and address gaps in available data through targeted data generation; 3) to support the analysis of these integrated data resources that link data on genetic and nongenetic factors (e.g., diagnostic, clinical, behavioral, neurophysiological, pharmaceutical and environmental exposures, complications of pregnancy and peri-natal events) to explore contributors to the causes of autism and/or to identify patterns associated with treatment/intervention outcomes and the use of services for autism; and 4) to provide a venue for replication of these analyses by independent teams to validate findings and increase transparency in the conduct of science.
1#footnote1 A valid data source refers to any information that is methodologically sound, ethically obtained, and suitable for drawing meaningful and reproducible conclusions pertaining to the questions at hand.
Background
About 1 in 31 children in the United States has been identified with autism, which affects approximately 3-4 males for every female (CDC, 2025). Core features of autism are difficulties with social communication, social interaction and restricted, repetitive patterns of behavior or interests. Autism can be reliably diagnosed by experienced healthcare specialists as early as age 2. It is an extremely heterogeneous condition with a variable clinical presentation and differing service and support needs. Family studies in autism have identified high heritability. Genetic studies of autism in tens of thousands of persons have revealed a complex genetic architecture, including associations with single-gene disorders or aneuploidies, rare de novo copy number variants, single nucleotide variants, and heritable common polygenic risks. Autism risk also increases with parental age. Nongenetic factors (e.g., environmental chemicals, medications, maternal health) also contribute to the development and expression of autism but in general these are less well characterized and likely many remain to be discovered. The prevalence of autism among 8-year-olds has increased most years since the Centers for Disease Control and Prevention (CDC) began tracking the condition in 2000, and occurs across all racial, ethnic, and socioeconomic groups. This increase in autism prevalence is likely due to a combination of factors. Studies sponsored by the National Institutes of Health (NIH) and other federal and non-governmental organizations have identified multiple factors that may be driving this increase including, but not limited to, changes in diagnostic factors, (e.g., diagnostic ascertainment, diagnostic criteria and/or practices), greater parent and practitioner awareness, and increased development and availability of services and supports that motivate families to pursue a diagnosis. The contribution of other factors that may play a role in the rising prevalence are not well studied or understood and include changes in the patterns and composition of prenatal/early life environmental chemical exposures (e.g., air pollution), parental health factors (e.g., obstetric complications, cardiometabolic diagnoses, maternal stress, and infection and immune alterations), and birth factors (e.g., preterm or very preterm birth). One or more of these factors may be influenced by genetics or interact with genetic susceptibility, reflecting gene-environment interactions (GxE factors). One mechanism by which these nongenetic factors may alter risk is by increasing or decreasing gene expression through epigenetic mechanisms. The totality of nongenetic factors over the life course is represented by the concept of the exposome.
Innovative new approaches that reflect an exposomics framework and that apply new analytic and statistical methods to data derived from human epidemiology and clinical settings are needed to advance the study of GxE interactions in autism. Observations of altered prenatal and early-life brain development and dysregulation of neuronal and immune function in people on the autism spectrum and in some non-human model systems can provide clues to identify causative factors. In addition, the male bias and sex-related differences in presentation and underlying biology may also be informative. Large datasets from NIH supported research with clinical data, such as neurophysiological (e.g., electroencephalography, evoked responses, magnetoencephalography, functional magnetic resonance imaging) or other biomarkers, can be leveraged for these purposes. The development of new approach methodologies (NAMs), including complex human-based in vitro models (e.g., brain organoids), has shown promise in addressing some of the knowledge gaps in autism etiology by enabling the experimental study of biologic response to potential neurodevelopmental toxicants or pharmacological treatments and their interaction with genetic susceptibility.
Despite improvements in autism screening and diagnosis, people on the autism spectrum have poor long-term outcomes resulting in higher health care utilization. People on the autism spectrum are dying much younger than expected, including by suicide, and are at increased risk of health problems (e.g., cardiovascular and metabolic conditions, epilepsy, sleep problems, pain management). Approximately three out of every four people on the autism spectrum have a co-occurring mental health condition, including anxiety, depression, ADHD, or substance use disorder, among others. Individuals on the autism spectrum will require different supports and interventions across the life span - from the point of first concern to later adulthood. Without treatments and supports, these problems negatively impact quality of life, educational and employment outcomes, and community participation. Significant advancements have been made in the development of novel interventions for autism. However, most of the intervention research has targeted early childhood, and research on autism in adolescence and adulthood is scarce. As the field of autism intervention research has included older individuals, outcomes have expanded to include mental health, educational and vocational functioning, self-determination, and quality of life. The rapid growth of this population has stretched the capacity of multiple service systems to deliver education, health and mental health services, and home and community-based care, all of which influence healthy outcomes and wellbeing. Pharmacologic treatment studies have also been conducted, although none have shown efficacy in affecting the core features of autism. There is limited understanding of how the complexity of genetic and nongenetic etiologies, together with variation in phenotype, is predictive of treatment response. There are few data available to determine the effectiveness of interventions that are accessible in standard community settings using existing workforces. This ‘research to practice’ gap is a major impediment to improving the lives of individuals on the autism spectrum.
Given the significant increase in prevalence of autism over the past 25 years, enhanced, cutting-edge research is urgently needed to identify and understand the full complement of factors contributing to this rise. Additionally, the Interagency Autism Coordinating Committee and the Autism Collaboration, Accountability, Research, Education, and Support Act of 2024 have highlighted the increased need for evidence-based treatments and interventions that have a more immediate impact on improving the lives of those living with autism. Identifying effective and scalable interventions across the lifespan addressing the spectrum of abilities, identifying predictors of treatment response, and improving our understanding of mechanisms of co-occurring conditions that could serve as treatment/service targets in future clinical trials are critical to advancing autism research. These areas could be catalyzed with progressive data science methodologies and rich, extant data resources.
Sources for Autism-Related Data
NIH and many other federal agencies have contributed to the development of rich data resources related to autism over at least the last three decades. Researchers responding to this ROA should aim to leverage existing data, such as those available through NIH-funded repositories. A non-exhaustive list of data sources can be found on this website (ADSI Data Resources page). Additionally, applicants may propose to pool large cohort data from existing federally funded datasets and/or use alternative valid data sources, which could include both public and private data. The proposed use of valid foreign datasets may also be permissible with appropriate data use agreements provided there are no foreign subawards. In addition, applicants may propose research that will enhance existing dataset/s by generating a limited amount of new data. Examples include but are not limited to the use of extant biospecimens to characterize metabolic or epigenetic markers, the mechanistic impacts of environmental chemical exposures in brain organoids or other human-centric models, generation of new environmental exposure estimates using geographic information system (GIS)-based modeling, or adding additional common outcome, implementation, or service measures across multiple on-going community-based clinical trials. This ROA will not support data generation in non-human animal models. However, there are many existing data resources that are derived from animal experimentation (e.g. toxicity data, functional data of autism risk genes) where comparable human data are unavailable. When combined with human data (clinical, epidemiological, observational, etc.), these non-human datasets can enrich the data aggregation, hypothesis testing, and exploratory analyses that will be supported through this initiative. All use of data within this initiative shall be compliant with privacy and confidentiality requirements, applicable federal and state laws and regulations, HHS and NIH policy, determinations of any involved Institutional Review Board, data use limitations from informed consent documentation, associated data use agreements, and data repository policies.
Source: NIH
