Angelini partners with Quiver on AI-driven discovery in genetic epilepsies

DEEs are a group of severe, early-onset epilepsies often linked to genetic variants, but translating those genetic diagnoses into targeted, mechanism-based therapies has remained difficult. Currently, most care is focused on seizure control rather than disease modification.

Why precision epilepsy remains difficult

Although many developmental epilepsies are now genetically characterized, moving from a DNA variant to a drug remains technically demanding. Mutations in ion channel genes can produce either gain- or loss-of-function effects, and even identical variants may lead to different clinical trajectories. That variability is what complicates the development of standardized therapeutic strategies.

In addition, seizures in these disorders often arise from broader network-level dysfunction and not a single disrupted pathway, making it difficult to predict whether modulating one molecular target will translate into clinical benefit. As a result, precision approaches in epilepsy have progressed more slowly than in some other genetically defined neurological conditions.

Angelini has already placed bets in this space. In 2025, it partnered with GRIN Therapeutics to develop and commercialize radiprodil outside North America, a selective NMDA receptor modulator being studied in children with rare genetic epilepsies linked to GRIN mutations. Those disorders are caused by alterations in genes encoding NMDA receptor subunits, which play a role in excitatory synaptic signaling in the brain.

Angelini also signed an option agreement with South Korea’s Sovargen covering SVG105, a preclinical antisense oligonucleotide programme targeting components of the mTOR signaling pathway, another pathway implicated in certain genetic epilepsies and neurodevelopmental disorders.

Where Quiver fits

The collaboration adds a discovery layer to Angelini’s existing epilepsy assets, aimed at clarifying how specific mutations alter neuronal signaling.

Quiver’s technology is built around human neuronal models and large-scale electrophysiological measurements. The company generates induced pluripotent stem cell (iPSC)-derived neurons carrying defined genetic variants and records how those variants alter neuronal firing and network activity. It then applies machine-learning methods to integrate those functional readouts with genomic and molecular data, to link specific mutations to reproducible cellular phenotypes.

In genetic epilepsies, the causative gene is often identified, but the downstream biological effect is less clear. Mutations in ion channels or synaptic proteins can shift neuronal excitability in different directions, and those differences are not always predictable from sequence data alone. By focusing on functional neuronal measurements, Quiver can generate data that could help prioritize targets earlier in the discovery process.

Under the agreement, Quiver will apply its platform to generate and analyze human neuronal models of selected genetic epilepsies, using high-throughput electrophysiological profiling to characterize how specific mutations alter cellular activity. The resulting datasets will be shared exclusively with Angelini during the research term.

Angelini will fund the research activities and retain the right to advance drug targets emerging from the collaboration. If it chooses to do so, Quiver would be eligible for milestone payments and royalties tied to future development and commercialization.

AI discovery remains rare in epilepsy

While partnerships in genetic epilepsies exist, few examples combine AI-based discovery with a specific focus on epilepsy. Angelini’s earlier collaboration with JCR Pharmaceuticals centered on applying JCR’s J-Brain Cargo blood-brain-barrier delivery technology to identify brain-penetrant biologics for epilepsy.

In contrast, most high-profile AI-driven alliances in neurological disease to date have been broader in scope or focused on other indications. For example, Danish central nervous system (CNS) specialist Lundbeck has partnered with Iambic Therapeutics to use its AI-powered drug discovery platform for neurological target discovery, though in migraine rather than epilepsy, and has also joined forces with a national AI supercomputing initiative to accelerate drug discovery for brain disorders more generally. However, epilepsy has so far been a less frequent focus of AI-driven discovery partnerships compared with broader CNS indications.