UC Berkeley

In the last few decades, computers have drastically enhanced scientist’s workflows. Herin, we describe how computers can be leveraged to better understand problems in bionanoscience. In chapter 2, we explore how classical machine learning models can predict which proteins would adsorb to the surface of carbon nanotube and polystyrene nanoparticles only using the protein sequence. We leverage large online datasets and protein structure prediction tools to build a robust protein feature database to enable prediction. In chapter 3, we develop a novel microscopy method to identify the density of Dopamine D2 autoreceptors. The method utilizes pharmaceutical and computational tools to identify changes in dopamine release activity. We then apply this novel method to mice that were sensitized with cocaine to identify changes in receptor density within the Nucleus accumbens.