Machine learning–assisted biosensors utilizing a set of biological polymers

A set of biological polymers, including amino acids, peptides, polysaccharides, proteins, colorimetric/fluorescent dyes, and so on, can collectively identify various biomolecules and environmental toxicants. Machine learning can analyze such high-dimensional data to identify and quantify the target molecules. While only a few works have been demonstrated for identifying bacterial species and per- and polyfluoroalkyl substances on paper microfluidic chips, it can be potentially expanded to detect other targets, including proteins, hormones, disease markers, endocrine-disrupting chemicals, and micro/nanoplastics. In addition, the method can be demonstrated on sensor array platforms, including stretchable electronics. This chapter summarizes what has been demonstrated and discusses future implementations and their limitations.