RIKEN–Seiko Collaboration Center
JPAdvancing Digital Liquid Biopsy
Liquid biopsy is a minimally invasive diagnostic approach that reads a person’s condition from body fluids such as saliva, blood, and cerebrospinal fluid. However, molecules that signal disease are present in very small amounts and are mixed among many similar molecules. We work to achieve both sensitivity and specificity through digital bioanalysis, which counts target molecules by confining them one at a time in tiny reaction chambers. By expanding the range of targets from nucleic acids to enzymes and extracellular vesicles, we are developing tests that detect conditions such as infectious diseases, cancer, and dementia earlier and more accurately.
SATORI: rapid, sensitive infectious disease testing without nucleic acid amplification
Common tests for viral and bacterial infections include PCR, which amplifies nucleic acids, and antigen tests, which detect proteins and other targets. PCR is sensitive but takes around an hour to return results, while antigen tests are simple and rapid but have limited sensitivity. Tests that combine high sensitivity with speed have long been sought.
Our SATORI method (CRISPR-based amplification-free digital RNA/DNA detection) combines target RNA and DNA recognition by CRISPR-Cas13 and Cas12 with femtoliter-sized microchamber arrays, allowing us to count individual targets without amplification. Once Cas13 or Cas12 recognizes a target, it successively cleaves nearby fluorescent reporters; confining the reaction in a microchamber makes the activity of a single molecule appear as a bright fluorescent spot. This principle achieved femtomolar (fM) sensitivity in under five minutes (Commun Biol, 2021; Anal Chem, 2026c).
We automated the entire workflow, from pretreatment to extract pathogen genes from samples through detection and quantification, and demonstrated that acute respiratory infections (COVID-19 and influenza) can be tested using clinical samples in just 10–20 minutes (Commun Biol, 2022; Cell Rep Phys Sci, 2026). This has advanced our research beyond proof of principle toward rapid testing technology that can be used in real clinical settings.
We are now working to expand the number of targets that can be tested at once, guided by the needs of clinical practice. By combining Cas13a and Cas13b, we can detect two types of RNA simultaneously in different colors, enabling concurrent testing for COVID-19 and influenza (Anal Chem, 2026b). We have also applied Direct-SATORI, which omits nucleic acid purification, to multiplex diagnosis of plant viruses (Anal Chem, 2023).
Applications extend beyond infectious diseases. In collaboration with the Laboratory of Racing Chemistry, we are developing tests to address the potential future misuse of mRNA-LNP medicines in sports doping. We built a fully automated platform that directly detects mRNA-LNP medicines in blood without extraction or amplification, achieving fM sensitivity in about ten minutes. The platform was demonstrated in both human and equine plasma (Anal Chem, 2026a).
SEAP: detecting cancer through the activity of individual enzymes
Blood contains a wide variety of enzymes, and changes in their activity reflect the body’s condition. Conventional measurements, however, average together enormous numbers of molecules—from millions to billions—making it impossible to distinguish enzymes with similar functions. SEAP (Single Enzyme Activity-based Protein Profiling) encapsulates enzymes individually in microreactors and reads their activity one molecule at a time, enabling the identification of enzyme isoforms.
Specifically, we are developing a method to identify individual enzyme molecules by optimizing substrate reactivity for each isoform and using fluorescent substrates that emit different colors. We read the unique reaction pattern of an enzyme confined in a microreactor—“which substrates it reacted with, and to what extent”—through multicolor fluorescence to determine its isoform.
We have demonstrated that diverse enzyme families—including phosphatases (Sci Adv, 2020; Cell Rep Phys Sci, 2026), esterases (Adv Sci, 2023), and peptidases and proteases (Cell Rep Methods, 2024)—can be identified and quantified one molecule at a time. Cosomil, Inc. was founded as a startup to bring this technology into practice, and is advancing its development toward early cancer diagnosis using digital liquid biopsy.
Digital SERS: detecting biomolecules beyond fluorescence and enabling multiplex analysis
Single-molecule digital counting often uses fluorescence to read out enzyme reactions, but distinguishing enzymes that catalyze similar reactions has been difficult. Raman scattering is highly sensitive to small differences in molecular structure and is therefore useful for identification. Its weak signal, however, has limited its use in single-molecule measurements.
We immobilized aggregates of silver nanoparticles at the bottom of microchambers. Their surface-enhancement effect amplified Raman scattering from the reaction products of individual enzymes by up to one million times. A fast wide-field Raman microscope that detects selected wavenumbers measured approximately 45,000 chambers in 8.5 minutes. For acetylcholinesterase, we achieved 10 fM sensitivity, approximately 100 times better than the conventional method.
Because differences in product structure appear as shifts in SERS peak wavenumbers, we can simultaneously identify and quantify acetylcholinesterase and butyrylcholinesterase, which could not previously be distinguished. In an evaluation using clinical samples, we found significantly fewer acetylcholinesterase molecules in the cerebrospinal fluid of patients with vascular dementia (ROC AUC = 0.8), indicating its potential use in dementia stratification (PNAS, 2025).
Expanding to new measurement targets
Liquid biopsy targets extend beyond nucleic acids and proteins. Extracellular vesicles, including exosomes, and small molecules such as metabolites are also important sources of information about the body’s condition. Their abundance and composition change in response to cellular activity and tissue changes, making them promising new indicators of disease onset and progression. By focusing on diverse molecules and particles, we aim to reveal biological information that cannot be obtained by analyzing nucleic acids and proteins alone.
We are also developing new methods to measure these targets one particle or molecule at a time. By detecting individual particles and molecules, we aim to identify and quantify differences that would be obscured by population averages, as well as components present in very small numbers. We are expanding the range of measurable molecules by designing reaction systems, devices, and detection and analysis methods optimized for each target. Furthermore, by combining different types of biological information, we aim to characterize disease from multiple angles and broaden the range of diseases that can be diagnosed.
