Quantum bioinformatics reality check: most studies never touch real quantum hardware

Quantum computing has been pitched as the future engine of computational biology, promising to crack problems that would take classical supercomputers centuries to solve. But a sweeping new systematic review suggests the field is far closer to its proof-of-concept childhood than its headline writers would have you believe. Researchers at Sapienza University of Rome and the University of Tuscia analyzed fifty-three experimental studies spanning genome assembly, protein folding, phylogenetics, and biomedical machine learning, and their verdict is stark: while quantum bioinformatics has produced viable prototypes, the vast majority of published results never touch real quantum hardware, and fewer than a third can be independently verified at all.

The review, published in Frontiers in High Performance Computing, was designed to separate theoretical promise from empirical reality. The authors systematically assessed experimental studies across four core biological domains, evaluating each one by its underlying computational paradigm, the scale of its datasets, the quality of its comparisons against classical baselines, and its adherence to open-science reproducibility standards. The findings reveal a field dominated by classical simulations and quantum-inspired heuristics rather than genuine quantum execution, with nearly 77 percent of implementations restricted to heavily reduced toy datasets due to current technological constraints.

The technical landscape of quantum bioinformatics is more fragmented than the term quantum computing suggests. The review distinguishes four paradigms: quantum annealing, gate-based quantum computing, quantum-inspired algorithms, and classical simulation. Quantum annealing, most commonly executed on D-Wave processors, encodes biological problems as Quadratic Unconstrained Binary Optimization (QUBO) problems or Ising models, where the solution corresponds to the minimum-energy configuration of a physical system. Gate-based quantum computing, by contrast, is a more general-purpose and theoretically universal model, manipulating qubits through sequences of quantum gates arranged in circuits. Quantum-inspired methods borrow concepts from quantum mechanics, such as superposition-like probability amplitudes, but execute entirely on classical hardware. Classical simulators, such as Qiskit, reproduce the behavior of quantum circuits without a physical quantum processor.

The distribution of these paradigms across the reviewed literature is telling. Classical simulations of quantum circuits represent the dominant approach, accounting for 47 percent of the 53 studies. Quantum annealing and quantum-inspired algorithms each account for 21 percent, while gate-based quantum computing, the paradigm most closely associated with universal quantum advantage, accounts for just 11 percent. Among studies utilizing real quantum hardware, D-Wave quantum annealers dominate the landscape, spanning three hardware generations from the D-Wave 2X through the 2000Q to the Advantage. Only six studies employed gate-based quantum processors, distributed across IBM, IonQ, Rigetti, and QuantumCTek systems.

Several factors explain D-Wave’s predominance. Many core bioinformatics problems, from genome assembly to RNA folding, codon optimization, and protein design, can be naturally reformulated as combinatorial optimization problems expressible as QUBO or Ising Hamiltonians, which is precisely the class of problems quantum annealers are purpose-built to solve. D-Wave systems also offer significantly higher qubit counts than current gate-based processors, with the Advantage2 providing over 4,400 qubits compared to the hundreds available on the largest gate-based devices. The D-Wave programming model is comparatively accessible: users formulate their problem as a QUBO matrix and submit it to a cloud-accessible processor without needing expertise in quantum circuit design or error mitigation strategies.

Gate-based quantum computers, by contrast, offer universal programmability and can in principle implement any quantum algorithm, including those with no natural QUBO formulation. The studies employing gate-based hardware exploited this universality for variational quantum eigensolver-based protein structure prediction, quantum circuit-based nucleotide identification, and quantum approximate optimization algorithm-based molecular pruning for protein docking. However, gate-based processors currently suffer from severe limitations in qubit count, gate fidelity, coherence times, and connectivity, which restrict practical problem sizes to toy-scale instances and require sophisticated error mitigation techniques. Consequently, gate-based approaches in bioinformatics remain largely confined to proof-of-concept demonstrations on small molecular fragments or simplified model systems.

The performance picture across the reviewed literature is uniformly positive but warrants cautious interpretation. Of the 53 studies, 40 include an explicit comparison between their quantum or quantum-inspired implementation and at least one classical baseline method. Strikingly, every one of these 40 studies reports performance that is either comparable to or better than the corresponding classical approach. Not a single study reports that a quantum implementation performed significantly worse than its classical counterpart on the tested problem instances. This pattern holds across all four thematic groups and all four quantum paradigms. However, the authors caution that publication bias, the well-documented tendency for studies reporting positive results to be published preferentially over those reporting negative findings, could play a significant role in this trend. It is implausible that no research group has ever observed a quantum implementation performing worse than a classical baseline on a bioinformatics task; rather, such negative results are unlikely to survive the publication process.

The dataset scale analysis reveals a paradox at the heart of the field. Of the 53 examined studies, 41, or 77 percent, were evaluated exclusively on toy or substantially reduced datasets, while only 12 employed datasets approaching real-life scale. The distribution across paradigms is highly asymmetric: among the 11 quantum annealing studies, 10 used toy or reduced datasets, and among the six gate-based studies, only one employed a dataset comparable to real-life cases. By contrast, among the 36 quantum-inspired or simulation-based studies, 10, or 28 percent, were evaluated on datasets of realistic scale, including applications in cancer genomics, graph classification, multi-omics classification, disease prediction, and medical imaging. This disparity directly reflects the hardware constraints of current quantum devices, creating a paradox in which the approaches most likely to demonstrate genuine quantum advantage are precisely those least able to be tested at a meaningful scale.

Reproducibility emerges as perhaps the most severe concern. Only 18 studies, or 34 percent, provide publicly available source code, while 35 do not share their implementations. Data availability is somewhat better, with 37 studies providing access to their experimental datasets. Critically, only 16 studies, or 30 percent, make both code and data simultaneously available, which is the minimum requirement for full computational reproducibility. At the other extreme, 14 studies provide neither code nor data, rendering their results entirely unverifiable by independent researchers. The challenges of quantum reproducibility are further complicated by restricted and expensive access to quantum processing units, temporal variability in hardware calibration that can change over hours to days, the probabilistic nature of quantum measurements, the black-box character of vendor platforms, and high sensitivity to hyperparameters such as annealing schedules, chain strengths, and variational circuit architectures.

The authors conclude that quantum bioinformatics has successfully produced viable proof-of-concept prototypes but is fundamentally limited by hardware scaling bottlenecks and a lack of standardized benchmarking. To move from conceptual demonstrations to practical scientific impact, they recommend that the community prioritize rigorous reproducibility practices, establish transparent evaluation metrics, and clearly distinguish genuine quantum execution from classical simulation. The field cannot mature, they argue, if the majority of published results remain unverifiable, and the path from prototype to practice will require interdisciplinary collaboration between quantum physicists, computer scientists, and domain experts in biology and medicine, along with community standards for benchmarking and transparent reporting that match the methodological expectations of both the quantum computing and bioinformatics communities.

Subject of Research: Beyond the hype: an empirical assessment of quantum computing in bioinformatics

Article Title: Beyond the hype: an empirical assessment of quantum computing in bioinformatics

Article References: Arcieri, M., Zuliani, P., Carotenuto, G., Carrus, M., & Castrignanò, T. (2026). Beyond the hype: an empirical assessment of quantum computing in bioinformatics. Frontiers in High Performance Computing, 4, Article 1923360. https://doi.org/10.3389/fhpcp.2026.1923360

Image Credits: AI Generated

DOI: 10.3389/fhpcp.2026.1923360

Keywords: quantum computing, bioinformatics, systematic review, reproducibility, NISQ, Beyond, hype, empirical, assessment, quantum, computing, scientific research