ӳ����ý

A new study of genes underlying neurodevelopmental variability has uncovered more than 70 genes that are strongly associated with autism, several for the first time, and hundreds of genes associated with more broadly defined neurodevelopmental conditions. The analysis is the largest of its kind to date and includes more than 150,000 participants, 20,000 of whom have been diagnosed with autism. The results offer the most comprehensive look yet at diverse forms of genetic variation within the protein-coding region of the genome in autism. The insights shed light on the molecular roots of brain development and neurodiversity, and provide new avenues for future research on the biology of autism.

The findings result from an international collaboration led by a team of scientists at the ӳ����ý of MIT and Harvard, including members of the Stanley Center for Psychiatric Research; Icahn School of Medicine at Mount Sinai; Massachusetts General Hospital; University of California San Francisco; and University of Pittsburgh School of Medicine. The datasets were derived from the Autism Sequencing Consortium (ASC), the Simons Foundation Powering Autism Research (SPARK) initiative, the Lundbeck Foundation Initiative for Integrative Psychiatric Research (iPSYCH), and the Center for Common Disease Genomics (CCDG) at the ӳ����ý of MIT and Harvard. The work is the culmination of an investment several years ago by these consortia to conduct large-scale genetic analysis for neurodevelopmental conditions and share these datasets on autism.

The new work appears in , alongside three related studies that use some of the same data to advance the understanding of the genetic basis of autism.

“Our discoveries were enabled by very large-scale, rich data collections in autism research and population genetic studies, as well as newly developed analysis methods that allowed us to explore the genetic roots of neurodevelopmental variability in new ways. In addition to the massive gene discovery efforts in the field, we are beginning to make inroads into understanding where, when, and how these genes exert their effects during neurodevelopment,” said co-senior author Michael Talkowski, institute member at the ӳ����ý and director of the Center for Genomic Medicine (CGM) at Massachusetts General Hospital.

“These collaborative efforts have yielded a wave of new genes and insights into neurodevelopment that can now be pursued in future studies, while sharing an analytic framework that can help uncover the role of diverse genetic variation in other traits and diagnoses,” said co-senior author Mark Daly, institute member at the ӳ����ý, founding chief of the Analytic and Translational Genetics Unit at Massachusetts General Hospital, and CGM faculty member.

"Genetic variants underlie neurodevelopmental outcomes,” said Stephan Sanders, a co-senior author from UCSF and member of the UCSF Weill Institute for Neurosciences. “That means finding numerous gene variants involved is critical to the field. This study shows the power of large-scale, cross-disciplinary research in human populations to accomplish that goal.”

Examining the exome

Early genetic studies examined chromosomal DNA under a microscope and used array-based methods to uncover duplications and losses of large DNA segments, known as copy number variants (CNVs), in autism. Recently, scientists have shown that smaller CNVs can be discovered by sequencing the whole genome or just its protein-coding portions, known as the exome. In this study, the team employed new methods co-developed with the ӳ����ý’s Data Sciences Platform to uncover smaller gene- and exon-level CNVs by analyzing exome sequencing data (GATK-gCNV). This allowed them, for the first time, to jointly process very large exome datasets using cloud-based computing to capture and analyze these changes in an integrated way with other types of genetic variation linked to autism, such as single-letter DNA changes and small bits of DNA that have been removed or added to the genome, called indels.

“The big step forward was our ability to analyze these kinds of variation together, at a large scale, in order to enhance our gene discovery power,” said co-first author Kyle Satterstrom, a computational biologist in the Daly lab.

To reveal genetic variants associated with autism, the researchers analyzed exome sequencing data from more than 60,000 people from SPARK, ASC, and the Simons Simplex Collection. They used a statistical framework developed by members of the team called “TADA” to integrate these data. The team identified 72 genes that are strongly associated with autism at the most stringent statistical threshold they applied, and as many as 185 genes with evidence of association at lower thresholds used. Most of the variants driving the gene discovery were newly arising (or “de novo”) in autistic individuals and were not observed in their parents, while a smaller but important contribution was derived from variants that were inherited.

By examining the relative contributions of different forms of genetic variation, the scientists were able to learn about the potential role of some of the 72 genes in giving rise to some features of autism. Most variants in the genes result in the loss of one copy of a gene, suggesting that they influence traits when a gene’s function is lost. Other variants change the spelling of a gene or are present in multiple copies, suggesting that they might play a role in autism by altering a gene’s function.

The researchers next combined data from the autism studies with a large dataset of 31,000 families in which a child was diagnosed with developmental delay and/or other neurodevelopmental conditions. “There’s so much overlap in the genes uncovered through studies of autism with those from studies of related conditions like developmental delay,” said co-first author Jack Fu, a postdoctoral fellow in the Talkowski lab. “One big task for our field is to disentangle where they’re distinct and where they overlap.”

These analyses revealed 373 genes associated with these neurodevelopmental outcomes and allowed the team to identify genes that are more associated with autism rather than other neurodevelopmental conditions, and vice versa. The subset of genes that are more strongly tied to autism also had some overlap with genes uncovered in an earlier study of schizophrenia, suggesting possible shared biological pathways underlying the two conditions.

Expression timing

In an analysis led by co-first author Minshi Peng, then a graduate student in the lab of co-senior author Kathryn Roeder, a professor of statistics and data sciences at Carnegie Mellon University, the scientists examined the expression, or activity levels, of the genes they uncovered in developing human neurons. They learned that genes linked predominantly to developmental delay tend to be active in early neuronal development, whereas autism-related genes tend to play a role in more mature neurons.

Researchers in the Stanley Center for Psychiatric Research and many others are following up on these findings by studying some of the genes to help reveal their function in neurodevelopment. “This is just the first discovery step,” said co-first author and ӳ����ý associate member Harrison Brand, an assistant professor in neurology at Massachusetts General Hospital and Harvard Medical School. “Now that we have this list of high-confidence genes, we can take it to the functional modeling stage to explore the biological mechanisms that underlie the features of autism.”

With new data expected soon from the SPARK consortium and plans to analyze whole-genome data that includes both coding and non-coding sequences, the scientists hope to fill out a more complete picture of the genetic architecture of autism and related conditions.

Co-senior authors of the study also include Joseph Buxbaum (Icahn School of Medicine at Mount Sinai) and Bernie Devlin (University of Pittsburgh School of Medicine).

Funding for this research was provided in part by the Simons Foundation for Autism Research Initiative, including the SSC-ASC Genomics Consortium and the SPARK project, the National Human Genome Research Institute, the National Institute of Mental Health, the National Institute of Child Health and Human Development, and the Seaver Foundation.