Single-walled carbon nanotubes (SWCNTs) emit photostable near-infrared (NIR-II) fluorescence that is ideal for optical sensing. Their properties, such as sensor brightness, selectivity, and sensitivity, can be controlled by wrapping their surfaces with biological and synthetic moieties. This presentation focuses on our advancements in bioengineering DNA, which self-assembles onto SWCNTs through the stacking of its bases on the SWCNT surface. Since different DNA sequences have unique effects on SWCNT fluorescence, DNA provides a nearly limitless sequence space for tuning the sensing properties of SWCNTs. However, the relationship between the DNA sequence and its effects on optical properties of SWCNTs is unknown.
We develop a technique inspired by directed evolution to engineer DNA sequences for optical SWCNT sensing. Through iterative rounds of screening, selection and DNA mutation,
we have engineered DNA-based optical SWCNT sensors for mycotoxins, glucose, neurotransmitters, nitric oxide (NO), and cancer biomarkers. By integrating machine learning, we enhance this process by predicting beneficial DNA mutations and designing diverse libraries to target analytes previously lacking responsive sequences. This strategy enables the automated development of analyte-specific sensors. It also unlocks new applications for optical DNA-SWCNT sensor arrays — an “optical nose” or “o-nose” — for versatile disease diagnostics based on pattern-based optical recognition.