Integration with AI

One of the most fundamental quantities in science is counting. For example, we count objects as “one” or “two.” However, since the discovery of Avogadro’s number, the sheer enormity of molecular numbers has made it practically impossible to count molecules one by one. At that time, there was no method to detect and count individual molecules.

Today, however, the concept of “one” has become critically important. A single cancer cell or a single protein molecule can have a profound impact on an entire system. This shift is not limited to science; just as a single unconventional individual can trigger transformative innovation in society, we are entering an era in which small units carry immense value. To precisely capture this “one,” we are developing a single-molecule counting method by combining nano-gap–based single-molecule measurement technology with AI. This approach enables the detection of individual molecules, identification of their types, and accurate counting—capabilities that were previously unattainable.

At present, we have successfully counted a single type of molecule, referred to as molecule A. In real samples, however, multiple molecular species—such as A and B—are usually present simultaneously. To address this challenge, we are developing AI-based techniques that can detect, identify, and count multiple types of molecules at the single-molecule level.

Once this technology is fully established, it will become possible to selectively detect, identify, and count target molecules with high precision from complex samples containing diverse substances. This eliminates the need for labor-intensive pre-processing steps required in conventional analytical methods, dramatically simplifying analytical workflows. Our research has the potential to fundamentally transform the field of analytical chemistry.