From Invisible Biology to Tomorrow’s Therapeutics
The Data No One Else Has
THE PROBLEMThe data have been fragmented, incomplete, and missing a critical layer.
Most drug discovery platforms are built on an incomplete picture of cancer biology — single modalities, different patient cohorts, and cell lines instead of real human tissue.
More compute applied to fragmented data does not solve the problem. It amplifies the mistakes already built into the foundation.
1965
Extrachromosomal circular DNA first documented in the scientific literature
0
Clinical platforms with access to this layer before Phinomics
100%
Of AI models, foundation models, and drug discovery platforms trained without extrachromosomal circular DNA data
Oncogenic drivers that pharma targets can exist in the genome, but be massively overrepresented in the extrachromosomal layer. Many cancers have hidden drivers that are missed entirely.
The Hidden LayerThe hidden 50% of cancer biology.
Extrachromosomal circular DNA — genetic elements outside the genome that drive cancer progression, drug resistance, and tumor evolution — have remained invisible to every existing platform.
Until now.
EvidenceFour Decades of a Known, Unsolved Problem
For more than 40 years, researchers have documented the clinical consequences of this invisible biological layer. Missed targets. Treatment resistance. Poor patient outcomes. The evidence was there. The ability to see it was not.
1983
Small Cell Lung Cancer
Double minutes drove reversible methotrexate resistance. Clinicians had no way to know the drug was already failing.
1984
Leukemia, Ovarian Cancer, AML
Extrachromosomal DHFR amplification drove relapse in multiple patients. Each treatment course amplified the resistance gene further. Standard karyotyping showed nothing.
1991–95
91% of amplified patient tumors carried only double minutes
Every genomic treatment decision of the era was made about a structure whose behavior was invisible and unaccounted for in any protocol.
2004
T-Cell ALL
5.6% of T-ALL patients carried an imatinib-sensitive oncogene on extrachromosomal episomes invisible to standard cytogenetics. All received chemotherapy instead of a drug that would have worked.
2014
Glioblastoma
The drug target disappeared onto extrachromosomal DNA under treatment and reappeared within 72 hours of withdrawal. Targeted therapy was aimed at a moving target no one could see.
2024
Pan-Cancer
Chemotherapy-pretreated patients had significantly more extrachromosomal DNA than untreated patients. Treatment selected for the resistance mechanism — undetected, in every case.
Extrachromosomal circular DNA is not one resistance mechanism sitting alongside the others. It's an upstream, systematic, and historically invisible failure mode that shows up at every stage where oncology drug development breaks down.
How It WorksThe first complete view of cancer biology.
Phinomics developed patented methods to capture circular DNA at 100× higher resolution than any existing sequencing approach, using matched tumor and normal patient samples. This is the Circulome: a first-in-class biological data layer that every other platform is missing.
We then integrate the Circulome with DNA, RNA, epigenetics, and proteomics from the same patient, creating a complete, multi-dimensional view of disease.
Explainable AI applied to complete biology reveals how disease actually works, not just what it looks like.
DEEP BIOLOGYThree properties no other dataset can claim simultaneously.
Phinomics combines a first-in-class Circulome data layer, true matched multi-modal integration, and matched normal tissue from every patient. Together, these create a biological foundation no other platform can replicate.
More compute on better data compounds in ways that more compute on fragmented single-layer observations simply cannot. Fragmented data amplifies mistakes while matched multi-omic data amplifies understanding.
A first-in-class data layer
The Circulome captures biology invisible to every competing platform, including extrachromosomal circular DNA at 100× higher resolution from real matched human samples.
True matched multi-modal integration
Every molecular layer comes from the same patient samples. Not aggregated across cohorts. Not inferred across experimental systems.
A matched normal tissue blueprint
Matched normal samples from every patient distinguish tumor-specific signals from background biology.
THE OUTPUTDeep Biology.
A multi-omic knowledge graph that reveals the causal mechanisms driving cancer — across biological layers that were previously invisible.
This unified system reveals the causal pathways driving cancer, identifying the exact vulnerabilities that can be targeted.
01
Novel therapeutic targets with defined modes of action
Target selection grounded in causal multi-omic evidence, not correlative single-layer signals.
02
Resistance mechanisms identified before the drug enters trials
Extrachromosomal amplification as a resistance driver — identified at the discovery stage, not at clinical failure.
03
Patient stratification grounded in multi-modal signatures
Biomarkers derived from matched, integrated molecular data — not single-modality proxies.
04
Defensible IP grounded in mechanistic discovery
First-in-class data layer generates first-in-class targets — with the mechanistic rationale to defend them.
The clinical infrastructure for circular DNA has never existed.
We’re building it.
For the first time, cancer can be understood as it actually behaves — revealing the mechanisms that drive progression, resistance, and response.