Mount Sinai study establishes framework for precision psychiatry using genetic biobank recall

Inside the BioMe biobank at Mount Sinai, researchers had something most psychiatric studies do not: a massive collection of genetic information paired with real medical records from a genuinely diverse population. They decided to test whether they could use that combination to do something practical—find people carrying rare genetic variants linked to serious brain disorders, then invite them back for deeper study.

The variants they were hunting for are called copy number variants, or CNVs. These are stretches of DNA that appear in duplicate or are missing entirely, and certain ones substantially raise the risk of autism spectrum disorder, intellectual disability, and schizophrenia. The researchers identified 335 people in the biobank carrying these high-risk variants, then reached out to them alongside two comparison groups: 217 people with schizophrenia who did not carry the variants, and 340 neurotypical controls. In total, they attempted to recruit 892 participants.

The response was modest. Eighteen percent of those contacted agreed to participate, and of those, only 8 percent completed the full battery of psychiatric and cognitive assessments the researchers had designed. That low completion rate might sound discouraging, but it was actually the point of the exercise. The team was not trying to prove that everyone would come back. They were trying to establish what actually happens when you attempt this kind of targeted recruitment in a real healthcare system—to create a working blueprint for others to follow.

What made this study distinctive was not just the genetic targeting, but the diversity of the people who showed up. The final cohort reflected BioMe's makeup: 37 percent identified as having African ancestry, 34 percent as Hispanic, and 26 percent as European. This matters because most psychiatric genetics research has historically skewed heavily toward European ancestry populations, leaving huge gaps in understanding how these variants affect other groups. Here, the researchers had a genuinely representative sample.

When the participants came in for their assessments, something important emerged. The detailed clinical evaluations—the face-to-face interviews, the cognitive testing, the structured psychiatric interviews—revealed characteristics that the electronic health records alone had never captured. The EHRs are useful for billing and basic documentation, but they miss the texture of a person's actual psychiatric and cognitive presentation. Direct assessment added dimensions that mattered for understanding how these genetic variants actually manifested in lived experience.

The study, published in npj Genomic Medicine, does not claim to have solved precision psychiatry. Instead, it offers something more grounded: a practical framework. Here is what worked. Here is what the response rates look like. Here is how to maintain diversity in your recruitment. Here is what you learn when you go beyond the medical record. Future researchers attempting similar work now have operational benchmarks—concrete numbers and methods they can use to design their own studies.

The implications point toward a different kind of psychiatry. If researchers can reliably identify people carrying genetic variants that substantially increase psychiatric risk, and if they can characterize those people in detail, then clinicians might eventually use similar approaches to stratify patients, to predict who might benefit from particular treatments, to catch developmental disorders earlier. The variants themselves do not determine destiny—many people carrying them never develop the associated disorders—but they are a signal worth understanding. This study shows that understanding them in diverse populations, at scale, within existing healthcare systems, is actually possible.