Machine Learning Approach to Bioengineering Optical Sensors Based on Carbon Nanotubes

Fluorescent nanomaterials benefit from tunable bandgaps and associated optical characteristics that are predicted by theory. Their optical characterizations, however, such as absorbance and fluorescence emissions, depend strongly on the environment of these nanomaterials. These measurements can thus vary based on differences in functional modifications, pH, saline concentration, suspension media, and other environmental factors. This environmental sensitivity serves as the basis of nanosensors that optically respond to changes in the nanomaterial’s environment. However, the design of these optical nanosensors is limited by the inability to predict the nanomaterial’s optical response (or lack thereof) to an environmental change.<br/><br/>This presentation discusses approaches inspired by synthetic biology for developing optical sensors. The presentation focuses on sensors based on the near-infrared fluorescence emissions of single-walled carbon nanotubes (SWCNTs). The SWCNTs are non-covalently functionalized with single-stranded DNA. The optical properties of these DNA-wrapped SWCNTs, including their responsivity and selectivity to different analytes, are tuned in a non-predictable manner by varying the DNA sequence. We develop an approach inspired by directed evolution, an iterative screening technique that is conventionally used to engineer proteins with unknown structure-function relationships, to tune the optical properties of these nanosensors in a guided manner [1-2]. The data collected from this screening approach are used to train algorithms to design diverse screening libraries [3]. These algorithms are further applied to predict DNA sequences that will yield the desired optical responses. Through computationally-assisted directed evolution, we have thus developed a versatile approach for engineering the next generation of rationally designed optical nanotube sensors for diverse applications.<br/><br/>[1] Lambert, B.; Gillen, A.J.; Schuergers, N.; Wu, S.J.; Boghossian, A.A. Directed evolution of the optoelectronic properties of synthetic nanomaterials, Chem. Commun. 55, 3239-3242 (2019).<br/><br/>[2] Lambert, B.; Taheri, A.; Wu, S.J.; Gillen, A.J.; Kashaninejad, M.; Boghossian, A.A. Directed evolution of nanosensors for the detection of mycotoxins, bioRxiv 2023.06.13.544576; doi: https://doi.org/10.1101/2023.06.13.544576 (2023).<br/><br/>[3] Rabbani, Y.; Behjati, S.; Lambert, B.P.; Sajjadi, S.H.; Shariaty-Niassar, M.; Boghossian, A.A. Prediction of mycotoxin response of DNA-wrapped nanotube sensor with machine learning, bioRxiv 2023.09.07.556334; doi: https://doi.org/10.1101/2023.09.07.556334 (2023).