Bipolar disorder has long been recognized as highly heritable, yet the biology underlying the condition has remained difficult to resolve. The genetics are complex, the clinical presentation is heterogeneous, and the field has historically lacked the scale of coordinated investment needed to connect discovery research with real-world care.
In the most recent episode of The Genetics Podcast, Dr. Patrick Short speaks with Dr. Cara Altimus, CEO of BD², and Dr. Ben Neale, Associate Professor at Harvard Medical School and Massachusetts General Hospital.
Together, they discuss how BD² is creating a new model for bipolar disorder research, one that brings together large-scale genetics, deep phenotyping, longitudinal clinical data, open data infrastructure, and a learning health network designed to move discoveries toward better diagnostics and treatments.
BD² began as a philanthropic initiative bringing families and funders together around a shared scientific agenda. In February 2026, it spun out as an independent nonprofit. Cara describes the mission as “breakthrough discoveries for thriving with bipolar disorder,” a phrase shaped directly by input from people living with the condition.
While bipolar disorder research has often focused on symptoms, diagnosis, and acute care, BD² is trying to build toward whole-person health. Their goal is to improve how people live with bipolar disorder across the full course of their lives.
The initiative now spans genetics, brain omics, neural circuits, clinical care, longitudinal data collection, and data infrastructure. Cara’s central point is that progress requires more than isolated discovery. It requires a coordinated system that can generate data, share it, interpret it, and feed insights back into care.
Ben describes the major genetic challenge in bipolar disorder. Common variant studies, especially through the Psychiatric Genomics Consortium, have identified many associations. These variants collectively explain meaningful genetic risk, yet each individual association tends to have a very small effect.
That makes them hard to interpret biologically. Many common variants are non-coding, difficult to map to specific genes, and challenging to model in the lab.
BD²’s genetics work is helping to push the field into rare variant discovery. Rare coding variants can have much larger effects, making them more actionable for biology. Ben explains that the current work has identified 13 distinct genes associated with bipolar disorder, with some variants shifting an individual’s estimated risk from around 1% to closer to 10%.
A central theme of the episode is the need to pair large-scale discovery with deeper clinical and biological characterization. BD²’s integrated network is designed to do this. Cara describes it as a longitudinal cohort study embedded within a learning health network. The cohort will follow individuals with bipolar disorder for five years, collecting brain scans, Fitbit data, blood-based assays, app-based surveys, wellness assessments, electronic health record data, and whole genome sequencing.
The network currently includes 12 sites across the United States and Canada, with nearly 1,500 participants. Samples from these participants are being sent for whole genome sequencing at the Broad Institute.
This structure allows BD² to ask a more translational question: when a person carries a specific genetic variant, what does that look like in real life? How does it relate to symptoms, mood patterns, cognition, treatment response, clinical trajectory, and day-to-day functioning?
For drug developers, the lesson is that genetic discovery only becomes clinically useful when it’s connected to phenotypic, longitudinal, and engagement infrastructure. Longitudinal phenotyping can help determine which biological signals matter clinically, which measures are most predictive, and which patient groups may be most relevant for future intervention studies.
Ben emphasizes that psychiatry still lacks a clear equivalent to the biomarker path seen in Alzheimer’s disease, where CSF and later blood-based markers helped connect biology, diagnosis, and drug development.
Cara notes that across BD² investigator meetings, different types of data are beginning to show measurable biological differences, from neuroimaging and immune profiling to cognition and wearable data. She describes this as an early but hopeful signal that bipolar disorder may have distinguishable biological profiles across multiple domains.
The long-term opportunity is to identify the most informative, predictive, and scalable data elements. BD² does not expect every future study to collect every possible data type from every participant. The goal is to learn which measures matter most, then expand those measures into broader populations and more geographies.
One of the most relevant parts of the conversation for biotech and clinical development is the discussion of how rare genetic findings could inform therapeutic strategy.
Ben draws an analogy from schizophrenia. Loss-of-function mutations in CHRM4, a muscarinic receptor gene, are emerging as a promising genetic signal. CHRM4 is also one of the targets of Cobenfy, a recently approved schizophrenia medication. For Ben, this illustrates how rare variant discoveries can point toward therapeutic hypotheses that may apply beyond the small number of patients carrying a specific mutation.
The initial trial strategy may be to start with people who carry the relevant variant, since their biology may be more directly connected to the therapeutic mechanism. If efficacy is observed there, researchers can then consider broader patient populations.
This is especially important in psychiatry, where trials can be long, objective readouts are limited, and symptom-based endpoints may not map cleanly onto underlying biology. Genetics and deep phenotyping may help de-risk development by identifying biologically enriched populations and more precise intervention strategies.
Cara also highlights BD²’s ambition to create a network capable of running clinical trials over time. That would allow discovery, characterization, and interventional research to operate within a connected infrastructure.
The episode also returns to the experience of people living with bipolar disorder. Cara explains that BD²’s original research priorities were shaped by patient input. Participants emphasized whole-life health and wellness rather than symptom control alone.
Diagnosis remains a major challenge. Cara notes that in the US, it takes an average of seven years to receive an accurate bipolar diagnosis. She also points to estimates suggesting that more than 50% of people living with bipolar disorder may not know they have the condition or may not have a diagnosis.
Ben adds that treatment goals in mental health are complicated by the fact that mood, motivation, behavior, creativity, and identity are deeply intertwined. Successful treatment should reduce distress and harmful symptoms while preserving as much of the person’s sense of self as possible.
This conversation shows how bipolar disorder research is entering a more integrated phase. Rare variant discovery is beginning to identify genes with stronger effects. Longitudinal cohorts are making it possible to connect those findings to clinical reality. Deep phenotyping may help define meaningful subtypes, biomarkers, and trial-ready populations.
For the field of precision psychiatry, BD² offers a practical example of what it takes to move from genetic association to biological understanding and toward better care.
Listen to the full episode below.