Live · Platform family · Vrtron Solutions
Genomic intelligence across living domains
One AI engine reading DNA across humans, animals and plants. Three vertical products built on a shared cross-kingdom substrate.
01 · Overview
GenomIx is not a single product. It is both the engine and the family of three vertical products built on it (Rx for humans, Vx for veterinary, Px for plant).
Most companies build a vertical first and try to extract the platform later. We did the opposite: built the engine first, then layered domain-specific knowledge on top. The result is that adding a new biological kingdom is a knowledge-layer change, not a re-architecture.
If the same engine works for humans, animals and plants, it is no longer "a veterinary tool" or "a clinical tool". It is a genomic intelligence engine that interprets DNA across biological kingdoms.
02 · Three products
Each product runs on its own domain with its own brand and knowledge layer. Underneath, they share the GenomIx engine.
GenomRx Interpret variants to identify drug-gene interactions, hereditary risks and clinically relevant genetic insights. Cross-references PharmGKB, ClinVar, gnomAD and CPIC guidelines.
Drug-gene interactions · Hereditary risks · Cancer-risk variants · Antimicrobial resistance
genomrx.com
GenomVx The same AI that reads human DNA also reads canine, feline, equine and bovine DNA. Drug sensitivity, hereditary conditions, breeding insights, production traits.
Drug safety · Hereditary conditions · Breeding insights · Production traits
genomvx.com
GenomPx Cross-referenced QTL databases and reference genomes (SL4, IWGSC, IRGSP, B73v5, more). Yield, climate resilience, pathogen resistance, trait discovery for major crops.
Yield · Climate resilience · Trait discovery · Pathogen resistance
genompx.com03 · The engine
The GenomIx engine is the genomic intelligence layer that powers Rx, Vx and Px. Built from the ground up around Evo 2 (a 7B DNA language model trained on 9.3 trillion nucleotides across 128,000+ species, ranked #1 globally for variant-effect prediction) and integrated with the Anthropic Claude API for clinical and agronomic narrative generation.
01
VCF, FASTA, gzipped or raw. Any sequencing platform or clinical lab.
02
Filter and normalise variants across reference genomes, with species-aware handling for animal and plant references.
03
Evo 2 (7B) scores every variant by functional impact, working across humans, animals, plants. Cross-kingdom from a single model.
04
Domain-specific curated databases bolt on per product: PharmGKB / ClinVar / gnomAD for Rx, OMIA + breed databases for Vx, QTL + trait literature for Px.
05
Claude generates the clinical / agronomic narrative; the engine produces a branded PDF report and a structured API payload.
06
Encrypted storage, audit trails, and a REST API so labs and downstream products can integrate without rebuilding the pipeline.
04 · Why this works
01
We built the engine before the products. Adding a new domain is a knowledge layer, not a rebuild.
02
Evo 2 was trained across 128,000+ species. It is learning general genomic structure, not just human medicine.
03
PharmGKB, ClinVar, OMIA, QTL databases — each product bolts on the right knowledge for its domain.
04
From VCF upload to interpreted report in under three minutes on our GPU hardware, including a branded PDF and a structured API payload.
For commercial partners, labs and clinical contexts: GenomIx isn't just three products. It is a platform you can build new domains on, with a substrate that already understands DNA across kingdoms.
