Beck, DavidVaughan, Joshua CPhillips, Christian2024-09-092024-09-092024-09-092024Phillips_washington_0250O_27037.pdfhttps://hdl.handle.net/1773/51830Thesis (Master's)--University of Washington, 2024Modern computational resources have revolutionized the way scientists understand the sequence-structure relationship. Combinations of AlphaFold2 predictions and bespoke machine learning models can generate variance in protein sequences targeting a desired characteristic. To understand and ground the success of these models, molecular dynamics simulations can be used to screen proposed mutants for desired characteristics and function. In this study, molecular dynamics simulations are used to validate outputs from NOMELT, a large language model targeting protein thermostability, and propose a novel, computationally designed thermostable red-emitting fluorescent protein. Demonstrated by established molecular dynamics campaigns used for assessing protein stability and thoughtful structural analysis for two use cases, NOMELT is capable of increasing the melting temperature of a given protein sequence while maintaining complex protein structure and functionapplication/pdfen-USnoneComputationFluorescent ProteinsMachine TranslationMolecular DynamicsProtein EngineeringThermostabilityChemistryChemical engineeringBiophysicsChemistryThesis