DNAstack
SNVs
…GGACTTATCGA…
…GGACCCATCGA…
Tertiary Analysis
Built for HiFi.
Wired into the World.
The first long-read tertiary analysis platform built to connect the largest global federated HiFi database on day one.
See it live at ESHG 2026
Bring your hardest case. Alex Robertson will walk you through a federated query across the HiFi Solves network, live.
Powering the world’s largest long-read clinical consortium
Why Omics AI
One platform, every variant class, the whole network
A single long-read genome captures every class of variant — and makes it instantly comparable across the network — for clinical and research teams alike.
Built for long-reads
Every variant class from a single HiFi genome: SNVs, indels, structural variants, repeat expansions, CNVs, paralogs and methylation (coming soon). All fully phased, on one reference.
4× more rare coding SVs than short-read
4,891
variants called
HPO-aware ranking
ACMG/AMP classification
PATSCN1A0.97
LPAMECP20.89
VUSCACNA1A0.62
6 on the shortlist, evidence attached
AI that shows its work.
HPO-aware ranking and ACMG/AMP classification narrow thousands of calls to a short, defensible list, with the evidence behind every score.
From 4,891 variants to a shortlist of 6
Federated by design
Ask one question and reach data held across many institutions. The compute travels to the data, the data never moves, and only the aggregated answer comes back.
Your data stays in your environment, always
Built on open standards
Omics AI is built on GA4GH standards for discovery, access and analysis, including Beacon, Data Connect, Passports and WES. The network interoperates with the tools you already use, with no vendor lock-in.
Open GA4GH standards, no lock-in
The Network
Your hardest case may already be solved.
Just somewhere else.
Every site that joins makes the next diagnosis more likely, for everyone. One query reaches the whole network, and no institution moves its data.
100%
of known clinically relevant variants detected by HiFi in the consortium's first major study
4x
more rare coding structural variants found than short-read whole genomes
15+
institutions across 11 countries, live on the HiFi Solves network today
How it works
From a genome to
an answer.
Three steps. Bring your own pipelines or use ours.
1
Bring your data
PacBio HiFi or Nanopore, from your own cloud. Nothing leaves your institution.
2
Analyze and prioritize
Ask one question and reach the world's largest long-read consortium. The compute travels to the data, the genome never moves, and only the answer comes back.
3
Reach the network
HPO-aware ranking and ACMG/AMP classification narrow thousands of calls to a short, defensible list, with the evidence behind every score.
Ready when you are
See what your unsolved cases look like on HiFi
Bring us your hardest case, solved or unsolved. We'll show you what Omics AI surfaces and what the consortium remembers.
30 minute walkthrough, no slides.