August 27, 2026 — Autism Spectrum Disorder (ASD) is a highly heritable neurodevelopmental syndrome having diverse clinical presentations, with profound autism being linked to hundreds of risk genes. For scientists this complexity makes it especially challenging to determine how different mutational variants can produce similar consequences in the developing brain.

In a pivotal study published in Science on August 27, scientists at the Quantitative Biosciences Institute (QBI) and Department of Psychiatry and Behavioral Sciences at the University of California San Francisco (UCSF) built a molecular road map linking 100 high confidence ASD risk genes through a surprisingly interconnected network of crucial protein interactions.

This collaborative effort, led by Belinda Wang, MD, PhD, Rasika Vartak, PhD, Kelsey M. Hennick, PhD, Yefim Zaltsman, PhD, Zun Zar Chi Niang, PhD, along with co-corresponding authors Kirsten Obernier, PhD, Tomasz J. Nowakowski, PhD, Matthew W. State, MD, PhD, A. Jeremy Willsey, PhD, and Nevan J. Krogan, PhD, brought together over 70 researchers, psychiatrists, clinicians, and computational experts in artificial intelligence, and represented a major milestone for the Psychiatric Cell Map Initiative (PCMI)—a partnership established over 10 years ago with the goal of identifying new therapeutic targets for neurodevelopmental and psychiatric conditions.

Genes provide the molecular instructions or ‘code’ for a specific protein product, and mutations in this code can have profound consequences on the shape of the protein, as well as where it goes and what it can physically interact with inside the cell. Protein interactions make up intricate functional pathways and protein complex machines that carry out cellular processes important for neuronal signaling and brain development. By mapping protein interaction networks for ASD risk genes and their mutations, researchers can identify key areas where multiple pathways converge on a shared ‘molecular hub’, similar to how drivers on different roads can be led to the same main highways.

In the recent publication authors report that the 100 ASD risk genes make over 1800 protein interaction connections—many of which were altered by autism-associated mutations—making this the largest ever experimentally derived molecular map for autism. Surprisingly, many of the ASD risk gene protein networks converged on a few functional pathways and protein complexes, with many patient-derived mutations either disrupting or rewiring new connections to those same shared pathways and complexes. Researchers validated key interactions in lab-grown neurons and brain organoids, finding that mutations that disrupted protein networks had direct consequences on neurodevelopment and neuronal functions.

Future research could allow scientists to develop new therapies that target these shared molecular hubs instead of individual mutations. “After the initial excitement of discovering rare mutations that cause common forms of autism, the reality of how hard it would be to develop medicines to target the most severe end of the autism spectrum became abundantly clear,” said Matthew W. State, MD, PhD, a senior author and Chair of the Department of Psychiatry and Behavioral Sciences at UCSF. “This current work opens up a whole new world of possibilities for therapeutic targets and promises a generation of novel drugs that can transform what we are able to do in the clinic.”

By collaborating with experts in AI-based protein interaction predictions, researchers were also able to model how these ASD-associated protein interactions look in 3D-space. Using these predictions allowed researchers to identify surfaces on the proteins that could be targeted by drugs.

“For the first time, we can point to the precise molecular mechanisms that cause autism– not just the genes associated with the disease, but the actual protein interactions that go wrong and the specific interfaces we can target with drugs,” said Nevan J. Krogan, PhD, Professor at UCSF, Director of QBI, Senior Investigator at Gladstone Institutes and senior author of the study. “We are not just finding targets, but getting deep insights that inform us how to target them ultimately. Importantly, the framework we have built here is not limited to autism. This is a blueprint for translating any genetic disease, including neurodegeneration and oncology, into a therapeutic strategy, and that is what we have been building QBI to do.”

Figure Explanation: Shown in the figure above is an overlay of two images blending into each other; on the left is a fluorescent microscopy image of lab-grown brain organoids and on the right is a network representation of the ASD risk gene protein interactions. In the microscopy image, the magenta-colored areas are what is known as ventricular zone (VZ)-like rosette rings. Surrounding the VZ-like rings are green-stained cortical neurons. This clear inner-to-outer arrangement reflects natural brain development. In the network diagram, dark grey filled diamonds represent the 100 high confidence ASD risk genes, while the colorful circles represent interacting proteins with different functions. Credit for brain organoid image: Kelsey M. Hennick.

Full List of Authors: Belinda Wang, MD, PhD; Rasika Vartak, PhD; Kelsey M. Hennick, PhD; Yefim Zaltsman, PhD; Zun Zar Chi Naing, PhD; Benjamin J. Polacco; PhD, Ali Bashir, PhD; Manon Eckhardt, PhD; Mehdi Bouhaddou, PhD; Jiewei Xu, PhD; Nawei Sun, PhD; Micaela C. Lasser, PhD; Yuan Zhou, MS; Justin McKetney, PhD; Keelan Z. Guiley, PhD; Pawel Gniewek, PhD; Una Chan; Naufa Amirani, MA; Owen Griffiths; Nishant Chadha; Reshmi Tognatta, PhD; Merve Cakir, PhD; Martin Gordon, MS; Prachi Khare, MS; Sam Drake; Vanessa Drury, MPH; David F. Burke, PhD; Silvano Gonzalez, MD; Sahar Alkhairy; Reuben Thomas, PhD; Stephanie Lam; Montana Morris; Ethel Bader; Mélanie Dos Santos; Anastassia V. Komarova, PhD; Maxwell Bennett; Craig Ennis, PhD; Octavio Castillo; Yvonne Lim; Robert Martin; Meghan Seyler, MEM; Tierney Baum, PhD; Rebecca Krasnoff; George Wang, ME; Sagnik Middya, PhD; Sheng Wang, PhD; Presley Pham; Juan Arbelaez; Dexter Pratt, PhD; Sofia Bali, PhD; Shivali Chag, MS; Julia A. Kaye, PhD; Nadir Mahmood, PhD; Lee Spraggon, PhD; Thomas Rolland, PhD; Shawn Hervey-Jumper, MD, FAANS; James S. Fraser, PhD; Thomas Bourgeron, PhD; Steven Finkbeiner, MD, PhD; Caroline Demeret, PhD; Danielle L. Swaney, PhD; Sourav Bandyopadhyay, PhD; Trey Ideker, PhD; Pedro Beltrao, PhD; Helen Rankin Willsey, PhD; Ruth Hüttenhain, PhD; Kirsten Obernier, PhD; Tomasz J. Nowakowski, PhD; Matthew W. State, MD, PhD; A. Jeremy Willsey, PhD; Nevan J. Krogan, PhD.

About the UCSF Department of Psychiatry and Behavioral Sciences: The UCSF Department of Psychiatry and Behavioral Sciences and the Langley Porter Psychiatric Institute are among the nation's foremost resources in the fields of child, adolescent, adult, and geriatric mental health. Together they constitute one of the largest departments in the UCSF School of Medicine and the UCSF Weill Institute for Neurosciences, with a focus on providing unparalleled patient care, conducting impactful research, training the next generation of behavioral health leaders, and expanding access, awareness, and advocacy across the field of behavioral health. Learn more at psychiatry.ucsf.edu. 

About the Quantitative Biosciences Institute (QBI): The Quantitative Biosciences Institute (QBI) is a University of California organized research unit reporting through the UCSF School of Pharmacy. QBI fosters collaborations across the biomedical and physical sciences to advance interdisciplinary approaches to human disease and therapeutic discovery. QBI’s disease-agnostic, uniquely integrated technological platforms have generated insights across cancer, neurodegeneration, and infectious diseases, leading to the formation of spinout companies, including Rezo Therapeutics. QBI incorporates the UCSF division of QB3, a multicampus UC institute that supports bioscience research and innovation in California. Learn more at qbi.ucsf.edu.