Functional genomics · AI drug discovery
Most cancer variants are still a mystery. We're changing that.
NeoVariant generates high-throughput functional evidence across the full cancer mutational spectrum — fusions, resistance mutations, and variants of unknown significance (VUS) — then layers AI to turn raw sequencing data into actionable biology. Our goal: no VUS left untreatable.
Every VUS — functionally characterized, every time
The problem
Sequencing has outpaced our ability to interpret what it finds
More than half of the variants detected in cancer genomes are still classified as uncertain — spanning the entire mutational landscape: point mutations, fusions, indels, and resistance variants. Oncologists are left without clear treatment guidance, and pharma without reliable biomarkers for trial design. The bottleneck isn't sequencing — it's functional evidence.
The platform
Functional evidence across the full cancer mutational spectrum
We generate large-scale functional data on cancer variants and layer AI on top — turning raw variant calls into driver classifications, drug-response predictions, and resistance forecasts. We started with oncogenic fusions because no one else has solved them at scale. We are built to go further.
-
01
Oncogenic fusions
Fusions are where NeoVariant starts. No existing platform maps fusion variant behavior — across gene partners, breakpoints, and drug pressure — at the scale we can.
-
02
Acquired resistance mutations
The same platform profiles the mutations that emerge after treatment — building a functional picture of resistance before it appears in patients, so pharma can design ahead of it.
-
03
The full mutational spectrum
Fusions are the beachhead. The platform is designed to cover point mutations, indels, and any class of VUS — because the vision is that no cancer variant should remain functionally uncharacterized or untreatable.
Each cell represents variant functionality
What we offer
Built for pharma and biotech — at every stage of the discovery pipeline
NeoVariant's platform is configurable to where your program needs it most. We work alongside your teams to generate the functional evidence your pipeline is built on.
Custom platform deployments
Every program is different. We configure the NeoVariant Platform to your variant set, target class, and indication — generating the specific functional dataset your program needs, not a generic one.
New target identification
Functional screening at scale surfaces oncogenic variants that observational sequencing alone misses. We identify which variants are true cancer drivers and which are noise — before your program commits to the wrong target.
Patient stratification
We profile drug response across a comprehensive catalog of variants, giving you the biomarker evidence to match the right patients to the right trials — and design inclusion criteria grounded in real functional biology.
Hit-to-lead optimization
Candidate molecules are screened against NeoVariant's full variant library, including primary targets and the resistance mutations most likely to emerge — so you optimize leads against the biology that actually matters in patients.
Team
Built by people who've done the underlying science
Priyanka Bajaj
Postdoctoral scholar, Bioengineering and Therapeutic Sciences, UCSF. PhD in Molecular Biophysics, Indian Institute of Science.
Logan Leak
Postdoctoral scholar, Department of Medicine, UCSF. PhD in Cancer Biology, Stanford. Fulbright Scholar 2026.
Advisors & Mentors
James Fraser
Professor & Chair, Bioengineering and Therapeutic Sciences, UCSF · QBI
Willow Coyote-Maestas
Assistant Professor, Bioengineering and Therapeutic Sciences, UCSF · Chan Zuckerberg Biohub · QBI
Randal Goomer
Principal Gen AI Lead, Amazon Web Services · Director Precision Healthcare, GE Healthcare · Founder, Avryll Therapeutics
Contact
No variant of unknown significance should remain untreatable.
We're building the platform to make that possible — and we're looking for pharma and biotech partners, investors, and collaborators who want to help.