Date of Award
Santa Clara : Santa Clara University, 2017.
The use of engineered nanomaterials (ENMs) in consumer and commercial products is increasing rapidly. The small size and high surface reactivity of ENMs gives them a range of attractive properties, and allows them to be incorporated into various materials. These properties make ENMs very appealing to modern industry, but also make ENMs toxic, causing serious health and environmental concerns. This toxicity is largely driven by the formation of a protein corona on the surface of ENMs. This protein corona is caused by proteins encountered in biological systems that bind to the surface of (ENMs). Despite the importance of the protein corona, little research has been done to control protein corona formation and model the biological conditions that contribute to protein-ENM binding. We present a quantitative characterization of proteins found bound on the surface of silver ENMs. A matrix of protein- ENM reactions were evaluated, including varied ENM sizes and surface coatings, as well as solution conditions (e.g. salt concentrations). Machine learning (random forest) classification was applied to this protein-ENM data matrix to evaluate the competing roles of the biophysical properties of proteins, ENM properties, and solution conditions in mediating formation of the ENM protein corona. The resulting model offers an accurate prediction of protein enrichment on ENMs with a receiver operating characteristic-score accuracy of 0.83. The effects of each variable in the formation of the ENM protein corona is calculated to provide recommendations for mechanistic models based upon protein quantification. Our model offers the framework to engineer ENMs to minimize the binding of toxic proteins, or maximize the binding of non-toxic proteins on the surface of ENMs.
Findlay, Matthew, "Machine Learning Offers Predictive Insight into the Silver Nanomaterial Protein Corona" (2017). Bioengineering Senior Theses. 63.