Japanese quantum algorithm solves metabolic problem

Ashish Joshi and Takahiko Koyama are behind the paper circulating in posts about a “Japanese quantum algorithm” for metabolism. The work is real: it was published in June 2026 in *Machine Learning with Applications* under the title “Quantum algorithm for metabolic network analysis,” according to Koyama’s Researchmap profile and article records. (researchmap.jp) The authors are affiliated with Keio University’s Human Biology-Microbiome-Quantum Research Center, or WPI-Bio2Q, in Tokyo. (biorxiv.org) The paper’s core claim is narrower than some social-media retellings. The authors say they built a quantum algorithm for analyzing metabolic networks by focusing on flux balance analysis, a standard optimization framework used to estimate how metabolites move through a cell’s reaction network. (researchmap.jp) They describe the work as the first application of quantum algorithms to metabolic pathway analysis. ### What problem are the researchers actually solving? Flux balance analysis is the target. The bioRxiv abstract says metabolic systems involve thousands of reactions and that analyzing them requires solving large mathematical problems that can become computationally prohibitive. (researchmap.jp) The Keio team says it reformulated that optimization task for execution on a quantum computer. Nature Protocols describes genome-scale stoichiometric modeling as a standard systems biology tool for modeling cellular physiology and growth, and says related methods are used to predict and design microbial communities. That is the biological backdrop for why faster or larger-scale optimization methods would matter. (biorxiv.org) ### What did the quantum method do differently? The authors say they used a quantum interior point method with a quantum subroutine for matrix inversion. In the abstract, they write that the approach uses quantum singular value transformation to make the metabolic optimization problem more suitable for a quantum computer. (biorxiv.org) The same abstract says the method offers a potential computational advantage over classical interior point methods for large, well-conditioned networks. That wording matters: the paper does not say it has already replaced classical metabolic modeling in practice. (nature.com) ### Did they simulate a whole cell or a microbial ecosystem? No. The paper says the authors demonstrated the method with numerical simulations on glycolysis and the tricarboxylic acid, or TCA, cycle. Those are core metabolic pathways, but they are much smaller than a full whole-cell model or a community-scale ecosystem model. (biorxiv.org) COMETS, a widely used microbial ecosystem platform described in *Nature Protocols*, can run simulations of multiple microbial species in spatially structured environments, and the protocol says such simulations can take from minutes to several days. That helps explain why researchers discussing future acceleration mention microbial communities, but that is a future use case rather than the benchmark reported in this paper. (biorxiv.org) ### How early is this in quantum biology? The authors themselves frame it as an early step. The abstract says the work “represents the first application of quantum algorithms to metabolic pathway analysis” and “pav[es] the way” for large-scale biological optimization. (biorxiv.org) A 2025 *Nature* review on quantum computing in precision medicine separately described whole-system biological modeling as part of the field’s long-term vision, not an already routine capability. Keio’s Bio2Q center also states that one of its goals is to develop methods to apply quantum computing to human biology. The paper fits that institutional agenda, but the reported result is still a pathway-level demonstration. (nature.com) ### What comes next from here? The next concrete reference point is the published article record itself. Sciety’s activity page shows the work first appeared on bioRxiv on October 27, 2025, and later as the journal article with DOI 10.1016/j.mlwa.2026.100913 on June 1, 2026. Future follow-up would likely come from the same Keio University groups — WPI-Bio2Q and the Sustainable Quantum Artificial Intelligence Center — if they extend the method to larger metabolic networks. (biorxiv.org) (sciety.org) (bio2q.keio.ac.jp)