A collaboration between the University of Oxford, the Wellcome Sanger Institute, and the University of Cambridge has achieved a milestone at the intersection of quantum computing and genomics: encoding a complete viral genome onto a quantum processor and running actual genomic analyses on it.
The team used IBM’s latest Quantum Heron r2 processor, a 156-qubit superconducting chip, to encode the full genome of Hepatitis D virus (HDV), the smallest known animal virus at approximately 1,700 nucleotides of circular RNA. The genome was represented across 117 qubits, achieving roughly an order of magnitude compression over prior quantum encoding efforts for genetic data.
The result, demonstrated as part of the Wellcome Leap Quantum for Bio (Q4Bio) Challenge, a $50 million competitive research program, represents a proof of principle that quantum computers can handle biologically meaningful genetic data, not merely toy sequences.
From letters to quantum states
Genomic data is, at its most basic level, a string of letters, A, C, G, T (or U for RNA viruses). Translating that string onto a quantum computer is non-trivial. Classical bits store either a 0 or a 1; qubits store superpositions of both, but the encoding scheme must be robust to noise and compatible with the limited connectivity of real quantum hardware.
The team converted the HDV genome into a Quadratic Unconstrained Binary Optimization (QUBO) formulation, a mathematical framework that maps naturally onto quantum annealing and variational quantum computing architectures. They then developed data partitioning and depth-reduction techniques to prepare the quantum states on IBM Heron’s superconducting chip, compensating for its limited coherence time and gate fidelity.
The encoding was verified through index-reported verification, a method that allowed the team to confirm the genome was correctly stored by checking that specific genetic positions could be retrieved on demand.
Actual genomic computation
Encoding a genome is one thing. Using it for science is another. The team went beyond storage to demonstrate four categories of real genomic computation on the quantum hardware:
1. Data encoding, converting raw nucleotide sequences into quantum states
2. Sequence alignment, mapping DNA/RNA fragments to a reference genome, the most common operation in bioinformatics
3. Pangenome assembly, constructing a composite genome from multiple individuals’ DNA, a computationally intensive task that scales poorly on classical computers
4. Phylogenetic tree construction, reconstructing evolutionary relationships between sequences
Each of these operations was executed in a hybrid classical-quantum workflow: the classical side handled problem formulation, iteration, and analysis, while the quantum processor was reserved for the most computationally challenging subproblems.
“Genomics is facing a data volume crisis,” said Dr. Sergii Strelchuk, Associate Professor at Oxford’s Department of Computer Science and the team lead, in an announcement from Jesus College Oxford. “Sequencing a single human genome produces roughly 200 gigabytes of raw data. As we enter the era of population-scale genomics, the computational demands are growing faster than Moore’s Law can address. Quantum computers offer a fundamentally different scaling path.”
The team
The collaboration brought together researchers from complementary domains. Prof. Richard Durbin, FRS (University of Cambridge Department of Genetics) provided the genomics expertise, Durbin was a central figure in the Human Genome Project and the 1000 Genomes Project. Dr. James McCafferty, Chief Information Officer at the Wellcome Sanger Institute, led the infrastructure side. The quantum methods drew on earlier theoretical work by Prof. Lloyd C. L. Hollenberg (University of Melbourne), who proposed the original framework for encoding genomic data in quantum states more than 25 years ago.
Additional team members included students Joshua Cudby and Orson Ye (Cambridge DAMTP), computational scientists Dave Holland, Dr. Peter Clapham, Robert Davies, James Bonfield, and Andrew Whitwham (Sanger Institute), and Floyd M. Creevey and Hitham T. Hassan (Oxford), whose earlier preprint (arXiv:2508.06184) developed the matrix product state encoding methodology used in this work.
The Q4Bio context
The result was one of six finalist demonstrations in the Q4Bio Challenge’s Phase III, a $50 million program funded by Wellcome Leap to catalyze quantum computing applications in biology and medicine. While the top prize in this phase was awarded to a different team (Algorithmiq / Cleveland Clinic / IBM for photodynamic therapy), the HDV encoding demonstration was widely noted as the clearest proof that whole-genome quantum analysis is technically achievable.
The team’s results were announced via institutional press releases in April and May 2026. Several research papers describing the methodology are under preparation.
Caveats
The demonstration is exactly that, a demonstration, not a production system. The HDV genome (1,700 nucleotides) is about 1.9 million times smaller than a human genome (3.2 billion base pairs). Scaling from 117 qubits to the millions of qubits that would be required for whole-human-genome quantum analysis faces formidable engineering challenges in error correction, qubit connectivity, and gate fidelity.
What the result does establish is that the quantum encoding of biologically meaningful genetic information, and the execution of real genomic algorithms on that encoded data, is no longer theoretical. The bridge between quantum information science and computational genomics has been crossed, even if the road ahead remains long.
This article is based on institutional announcements from the University of Oxford Department of Computer Science (May 22, 2026), the Wellcome Sanger Institute, Jesus College Oxford, and IBM Research, describing results from the Q4Bio Challenge. As of publication, the specific HDV encoding result has not yet appeared in a peer-reviewed journal. The underlying methodology is described in Creevey, F. M. et al., “Scalable Quantum State Preparation for Encoding Genomic Data with Matrix Product States,” arXiv:2508.06184 (2025).